# Import the needed modules 
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import os.path as op
import padasip as pa
import dataframe_image as dfi
import mne
import mne_nirs
import pywt
import scipy
import dabest
from itertools import compress
from mne.preprocessing.nirs import scalp_coupling_index, optical_density, temporal_derivative_distribution_repair, beer_lambert_law, _validate_nirs_info, source_detector_distances
from mne.preprocessing.nirs._beer_lambert_law import _load_absorption
from mne_nirs.preprocessing import peak_power, scalp_coupling_index_windowed, quantify_mayer_fooof
from mne_nirs.visualisation import plot_timechannel_quality_metric
from mne_nirs.channels import get_short_channels, get_long_channels, picks_pair_to_idx
from mne_nirs.signal_enhancement import short_channel_regression, enhance_negative_correlation
from mne_nirs.experimental_design import make_first_level_design_matrix
from mne.preprocessing import ICA, corrmap
from mne._fiff.constants import FIFF
from mne.filter import create_filter
from scipy.stats import norm, ttest_1samp, kurtosis, pearsonr, multivariate_normal
from scipy.signal import wiener, lfilter
from scipy.interpolate import CubicSpline, interp1d
from scipy.linalg import toeplitz
from fooof import FOOOF
from csaps import csaps
from sklearn.decomposition import PCA, FastICA
from sklearn.preprocessing import StandardScaler

C:\ProgramData\anaconda3\Lib\site-packages\paramiko\transport.py:219: CryptographyDeprecationWarning: Blowfish has been deprecated and will be removed in a future release
  "class": algorithms.Blowfish,
C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\2040918015.py:29: DeprecationWarning: 
The `fooof` package is being deprecated and replaced by the `specparam` (spectral parameterization) package.
This version of `fooof` (1.1) is fully functional, but will not be further updated.
New projects are recommended to update to using `specparam` (see Changelog for details).
  from fooof import FOOOF

# Write own fucntions
def reorder(raw_initial):
    """
    Input: raw object
    Function: reorders the channels to match the MNE standard
    """
    raw = raw_initial.copy()
    channel_names = raw.ch_names
    channel_names_ordered = []
    for i in range(0, int(len(channel_names)/2)):
        channel_names_ordered.append(channel_names[i])
        channel_names_ordered.append(channel_names[(i + int(len(channel_names)/2))])
    raw_ordered = raw.reorder_channels(channel_names_ordered)
    return raw_ordered

def add_info(raw_intensity, first_name, last_name, age, sex, EHI, annotations, duration_annotation, duration_rest, N_control):
    """
    Input: raw object, first name, last name, age, sex, EHI, annotations, duration of annotation, duration of baseline period, # of control trials
    Function: Returns raw object with added subject info and random allocated control trials within the baseline period
    """
    raw = raw_intensity.copy()
    
    # Add subject info
    raw.info["subject_info"] = {'first_name' : first_name, 'last_name' : last_name, 'age' : age, 'sex' : sex, 'EHI' : EHI}
    
    # attribute meaningful names to trigger codes. Include information about the duration of each stimulus (10 seconds). 
    raw.annotations.set_durations(duration_annotation)
    raw.annotations.rename(annotations)
    
    events, event_dict = mne.events_from_annotations(raw)
    
    # Define begin and end of rest period
    begin_rest = np.ceil(events[np.where(events[:,2] == event_dict['Baseline'])[0][0],0]/raw.info['sfreq'])
    end_rest = begin_rest + duration_rest - duration_annotation 
    
    # Remove start indicator of rest period
    raw.annotations.delete(np.nonzero(raw.annotations.description == "Baseline"))
    
    # Define indices of random selected control trials
    indices = np.random.choice(np.arange(begin_rest, end_rest), N_control, replace = False).astype(int)
    raw.annotations.append(indices, [duration_annotation]*N_control, ["Baseline"]*N_control)
    
    return raw

def crop(raw_input, t_before = 10, t_after = 60):
    """
    Input: raw object, time interval length before first event and time interval length after last event that should be included in the recording
    Function: crops raw object to a recording starting 't_before' s (default 10 s) before the first annotated event and ending 't_after' s (default 60 s) after the last event 
    """
    raw = raw_input.copy()
    end_rec = raw.get_data().shape[1]/raw.info['sfreq']
    t_start = events[0,0]/raw.info['sfreq'] - t_before
    t_end = events[-1,0]/raw.info['sfreq'] + t_after
    if t_start < 10:
        t_start = 0
    if t_end > end_rec:
        t_end = end_rec
    raw.crop(t_start, t_end)
    return raw

def get_indices(array, target_array):
    """
    Input: complete array of channel names, array of channels names to be removed
    Function: gives back indices of desired channels
    """
    lst = list(array)
    indices = list(np.arange(0, len(array)))
    for target in target_array:
        if target in lst:
            indices.remove(lst.index(target))
    return indices

def ch_names_L_R(raw):
    """
    Input: raw object
    Function: Separate list of channel names in 2 lists for the left and right hemisphere respectively
    """
    left, right = [], []
    for ch in raw.ch_names:
        if (int(ch[1]) % 2) == 0:
            right.append(ch)
        else:
            left.append(ch)
    return left, right

def scale_up_spectra(spectra, freqs):
    """
    Input: spectra, freqs
    Function: FOOOF requires the frequency values to be higher than the fNIRS data permits, 
    so we scale the values up by 10 here, and then will scale
    the frequency values down by 10 later.
    """
    freqs = freqs * 10
    return spectra, freqs

def get_IMU_data(filename, plot = False, sfreq = None, export = False, logy = True):
    """
    Extract IMU data from text file.
    Input: filename with .txt, plot (bool)
    Output: Three dataframes containing the complete IMU data, the accelerometer data and the gyroscope data respectively
    """
    
    # Read .txt file and convert to dataframe
    df = pd.read_csv(filename, sep=';', header=None, names=["time", "battery", "channels", "gyroX", "gyroY", "gyroZ", "accX", "accY", "accZ", "marker", "_"])

    # Select IMU data
    data = df.drop(0)
    markers = data["marker"].astype(float)
    IMU_data = data.drop(columns =['time', 'battery', 'channels', 'marker', '_']).astype(float)

    gyro_data = IMU_data.drop(columns = ['accX', 'accY', 'accZ'])
    acc_data = IMU_data.drop(columns = ["gyroX", "gyroY", "gyroZ"])
    
    if plot: 
        t = np.arange(gyro_data.to_numpy().shape[0])/sfreq
        gyro_data['time'] = t
        acc_data['time'] = t
        plot1 = gyro_data.plot(x = 'time', title='Gyroscope data', grid = True, xlabel = 'time (s)', ylabel = 'dps', logy = logy, figsize=(11,5))
        plot2 = acc_data.plot(x = 'time', title='Accelerometer data', grid = True, xlabel = 'time (s)', ylabel = 'm/s^2', logy = logy, figsize=(11,5))

        if export:
            fig1 = plot1.get_figure()
            fig2 = plot2.get_figure()
            fig1.savefig('Gyroscope_data.png')
            fig2.savefig('Accelerometer_data.png')
        
    return IMU_data, acc_data, gyro_data

def set_stimuli(raw_intensity, stimDur, restDur, t_begin, baseDur, t_rest, N_stimuli, N_control):
    raw = raw_intensity.copy()
    
    # Include information about the duration of each stimulus (seconds). 
    raw.annotations.set_durations(stimDur)
    
    # Define indices of random selected control trials
    end_rest = t_begin + baseDur - stimDur
    indices = np.random.choice(np.arange(t_begin, end_rest), N_control, replace = False).astype(int)
    raw.annotations.append(indices, [stimDur]*N_control, ["Baseline"]*N_control)

    # Define indices of hand grasping movement
    t_start = t_begin + baseDur + t_rest
    indices = t_start + np.arange(N_stimuli)*(stimDur + restDur)
    raw.annotations.append(indices, [stimDur]*N_stimuli, ["Task"]*N_stimuli)

    # Add info
    raw.info["subject_info"] = {'first_name' : 'Synthetic', 'last_name' : 'Data'}
    
    return raw

def find_duplicates_with_set(arr):
    seen = set()
    duplicates = set()
    for item in arr:
        if item in seen:
            duplicates.add(item)
        else:
            seen.add(item)
    return list(duplicates)

def MSD(x, k):
    w = 2*k+1
    N = len(x)
    msd = np.zeros_like(x)
    msd[:k] = np.NAN
    msd[-k:] = np.NAN
    for t in range(k,N-k):
        x_window = x[t-k:t+k+1]
        msd[t] = np.sqrt(np.sum(x_window**2)-np.sum(x_window)**2/w)/w
    return msd

# Set plot characteristics
%matplotlib inline
sns.set_theme() # nicer plots
# Load external plots?
plot = False
# export figures?
export = False

filename = 'MA_experiment_data.snirf'

# Read snirf data with motion artefacts
raw = mne.io.read_raw_snirf(filename, optode_frame='unknown', preload=False, verbose=None)

Loading C:\Users\fabia\fNIRS data analysis\Notebooks Master Thesis - Fabian Duren\Motion artefact removal selection\MA_experiment_data.snirf

# Reorder channels (Due to non-uniformity between mne_nirs and .snirf data)
raw_intensity = reorder(raw)

# Use self-written function to add info about the subject and the recording
annotations = {"1": "practice/Vertical", "2": "practice/Horizontal", "3": "practice/Random", "4": "test/Vertical", "5": "test/Horizontal", "6": "test/Random", "7": "Baseline"}
raw_intensity = add_info(raw_intensity, 'Subject', '1', 22, 'female', 'left', annotations, 3, 60, 10)

Used Annotations descriptions: ['Baseline', 'practice/Horizontal', 'practice/Random', 'practice/Vertical', 'test/Horizontal', 'test/Random', 'test/Vertical']

# Look at characteristics
raw_intensity

# Look at segments
events, event_dict = mne.events_from_annotations(raw_intensity)
plt.rcParams["figure.figsize"] = (10, 6) # (w, h)
fig = mne.viz.plot_events(events, event_id=event_dict, sfreq=raw_intensity.info['sfreq'])
if export:
    fig.savefig('events_MA.png')

Used Annotations descriptions: ['Baseline', 'practice/Horizontal', 'practice/Random', 'practice/Vertical', 'test/Horizontal', 'test/Random', 'test/Vertical']

# Get IDs of events
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
# Determine inter-stimuli intervals
inter_stim = np.diff(events[np.isin(events[:,2], ids)][:,0])/raw_intensity.info['sfreq'] - 3
if plot:
    # plot
    plt.hist(inter_stim)
    plt.xlabel('time (s)')
    plt.ylabel('Count')
    plt.title('Histogram of inter-stimulus intervals of MA experiment')
    if export:
        plt.savefig('Histogram_inter_stimuli_MA.png')

# As a table
raw_intensity.annotations.to_data_frame()

# Plot channels in 2D using build-in function of mne_nirs
sns.reset_defaults() # Turn of sns to be able to also plot contour of head
# dots = Center of channels
%matplotlib inline
raw_intensity.plot_sensors();

sns.set_theme()

if plot:
    # Plot channels in 3D
    subjects_dir = mne.datasets.sample.data_path() / "subjects"
    mne.datasets.fetch_fsaverage(subjects_dir=subjects_dir)

    brain = mne.viz.Brain("fsaverage", subjects_dir=subjects_dir, alpha=0.5, cortex="low_contrast")
    brain.add_head()
    
    identity_trans = mne.transforms.Transform("head", "mri") #transformed coord
    brain.add_sensors(raw_intensity.info, trans=identity_trans, fnirs=["channels", "pairs", "sources", "detectors"])

    # Add ROI: BA 1, 2, 3 - Primary somatosensory cortex (postcentral gyrus)
            # BA 4 - Primary motor cortex (precentral gyrus)
            # BA 5 - Somatosensory association cortex (NOT PRESENT IN LABELS)
            # BA 6	Premotor and supplementary motor cortex
    brain.add_label("BA1", hemi = "lh", borders=False, color='red')
    brain.add_label("BA1", hemi = "rh", borders=False, color='red')
    brain.add_label("BA2", hemi = "rh", borders=False, color='orange')
    brain.add_label("BA2", hemi = "lh", borders=False, color='orange')
    brain.add_label("BA3a", hemi = "rh", borders=False, color='yellow')
    brain.add_label("BA3b", hemi = "rh", borders=False, color='yellow')
    brain.add_label("BA3a", hemi = "lh", borders=False, color='yellow')
    brain.add_label("BA3b", hemi = "lh", borders=False, color='yellow')
    brain.add_label("BA4a", hemi = "lh", borders=False, color='green')
    brain.add_label("BA4a", hemi = "rh", borders=False, color='green')
    brain.add_label("BA4p", hemi = "lh", borders=False, color='green')
    brain.add_label("BA4p", hemi = "rh", borders=False, color='green')
    brain.add_label("BA6", hemi = "lh", borders=False, color='blue')
    brain.add_label("BA6", hemi = "rh", borders=False, color='blue')
    brain.show_view(azimuth=90, elevation=90, distance=500)

# Plot raw data with build-in plot function of mne_nirs
if plot:
    %matplotlib qt
    raw_intensity.plot(duration=60, show_scrollbars = True, clipping = None);

%matplotlib inline
fig = raw_intensity.copy().pick(np.arange(18,26)).crop(750,810).plot(duration=60)
if export:
    fig.savefig('MA_VI_1.png')

Using matplotlib as 2D backend.

# Convert raw intensity data to OD with build-in function of mne_nirs
raw_od = optical_density(raw_intensity)

Reading 0 ... 5162  =      0.000 ...   877.540 secs...

# Plot OD
if plot:
    %matplotlib qt
    raw_od.plot(duration=120, show_scrollbars = True, clipping = None);

%matplotlib inline
fig = raw_od.copy().pick(['S10_D2 760', 'S10_D2 850', 'S1_D7 760', 'S1_D7 850', 'S2_D8 760', 'S2_D8 850']).crop(350,410).plot(duration=60)
if export:
    fig.savefig('MA_VI_2.png')

# Convert OD to haemoglobin concentration with build-in function of mne_nirs
# Constant ppf value: PPF = DPF/PVC
raw_haemo = beer_lambert_law(raw_od, ppf=6) # ppf=6 is more inline with the community expectations and makes our results easier compared to the majority of the existing literature.

# Plot concentrations
if plot:
    %matplotlib qt
    raw_haemo.plot(duration=60, show_scrollbars = True, clipping = None);

%matplotlib inline
fig = raw_haemo.copy().pick(['S10_D2 hbo', 'S10_D2 hbr', 'S1_D7 hbo', 'S1_D7 hbr', 'S2_D8 hbo', 'S2_D8 hbr']).crop(350,410).plot(duration=60)
if export:
    fig.savefig('MA_VI_3.png')

# Calculate SCI with build-in function of mne_nirs
sci = scalp_coupling_index(raw_od, l_freq=0.7, h_freq=1.5, l_trans_bandwidth=0.3, h_trans_bandwidth=0.3, verbose=False)
# SCI bandpasses the signal with a FIR filter: l_freq=0.7, h_freq=1.5, l_trans_bandwidth=0.3, h_trans_bandwidth=0.3

# Plot histogram of SCI values
%matplotlib inline
fig, ax = plt.subplots(figsize=(12,6))
ax.hist(sci)
ax.set(xlabel='Scalp Coupling Index', ylabel='Count', xlim=[0, 1])
plt.vlines(0.75, ymin = 0, ymax = 18, color = 'r', label = 'SCI threshold')
plt.legend()
if export:
    plt.savefig('MA_SCI.png')
plt.show()

# Copy raw object to apply SCI 
raw_od_SCI = raw_od.copy()
# Mark channels with SCI < 0.75 as BAD
raw_od_SCI.info['bads'] = list(compress(raw_od.ch_names, sci < 0.75)) # 0.75 RECOMMENDED BY Pollonini et al. (2013)
print('The ' + str(len(raw_od_SCI.info['bads'])) + ' bad channels are: ' + str(raw_od_SCI.info['bads']))

The 28 bad channels are: ['S1_D1 760', 'S1_D1 850', 'S3_D1 760', 'S3_D1 850', 'S4_D2 760', 'S4_D2 850', 'S5_D1 760', 'S5_D1 850', 'S6_D2 760', 'S6_D2 850', 'S7_D1 760', 'S7_D1 850', 'S5_D3 760', 'S5_D3 850', 'S3_D5 760', 'S3_D5 850', 'S4_D6 760', 'S4_D6 850', 'S6_D6 760', 'S6_D6 850', 'S1_D9 760', 'S1_D9 850', 'S2_D10 760', 'S2_D10 850', 'S7_D9 760', 'S7_D9 850', 'S8_D10 760', 'S8_D10 850']

# Calculate windowed SCI with windows of 30 s
_, scores, times = scalp_coupling_index_windowed(raw_od_SCI, time_window=30)

%matplotlib inline
# Plot
plot_timechannel_quality_metric(raw_od_SCI, scores, times, threshold=0.75, title="Scalp Coupling Index Quality Evaluation");

# Bad channels are now displayed in red
%matplotlib inline
fig = raw_od_SCI.plot_sensors()
if export:
    fig.savefig('MA_SCI_topomap.png')

# View instances where a subset of channels may be contaminated by artifacts for a short duration of the recording
raw_od_SCI_PSP, scores, times = peak_power(raw_od_SCI, time_window=10)

# Plot
%matplotlib inline
fig = plot_timechannel_quality_metric(raw_od_SCI_PSP, scores, times, threshold=0.1, title="Peak Power Quality Evaluation") #see PHOEBE
if export:
    fig.savefig('MA_PSP.png')

sfreq = raw_od.info['sfreq']
IMU_data, acc_df, gyro_df = get_IMU_data(filename[:-5] + 'txt', plot = plot, sfreq = sfreq, export = export, logy =False)

# Create experimental design (recreate simple event-related experiment)
t_begin = 30 # 30 s resting state
baseDur = 60 # 1 minute baseline recording
t_rest = 30 # 30 s resting state
t_end = 30 # 30 s resting state
stimDur = 12
restDur = 23
N_stimuli =  10 # Create 10 hand grasping movements
N_control = 10 # Create 10 control segments
sfreq = 5.88 # Hz
totLength = t_begin + baseDur + t_rest + N_stimuli*(stimDur + restDur) + t_end # duration recording
t = np.arange(0, totLength, 1/sfreq)
tot_n_samples = len(t)
nchan = 4
nsignal = 2*nchan # HbO and HbR signal
ppf = 6
# Make 10 different synthetic datasets
n_set = 10

print('The total length of the recording is ' + str(totLength) + ' seconds')

The total length of the recording is 500 seconds

# Create empty raw object
data = pd.DataFrame(np.zeros((nsignal, tot_n_samples))) # 4 channel (4 hbo and 4 hbr signal)

ch_names = ['S1_D1 hbo', 'S1_D1 hbr', 'S3_D1 hbo', 'S3_D1 hbr', 'S5_D1 hbo', 'S5_D1 hbr', 'S7_D1 hbo', 'S7_D1 hbr']
ch_types = ['hbo', 'hbr', 'hbo', 'hbr', 'hbo', 'hbr', 'hbo', 'hbr']

info = mne.create_info(ch_names=ch_names, ch_types=ch_types, sfreq=sfreq)
raw_synths = dict()
for i in range(n_set):
    raw_synths[i] = mne.io.RawArray(data, info, verbose=True)

Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.
Creating RawArray with float64 data, n_channels=8, n_times=2940
    Range : 0 ... 2939 =      0.000 ...   499.830 secs
Ready.

# Set stimuli and optode locations
for i in range(n_set):
    raw_synth = set_stimuli(raw_synths[i], stimDur, restDur, t_begin, baseDur, t_rest, N_stimuli, N_control)
    raw_synth.info['chs'][0]['loc'] = np.array([-5.8205e-02, -2.2250e-03,  6.9370e-02, -5.1050e-02,  7.1800e-03, 7.4380e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][1]['loc'] = np.array([-5.8205e-02, -2.2250e-03,  6.9370e-02, -5.1050e-02,  7.1800e-03, 7.4380e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  8.5000e+02, np.nan, np.nan])
    raw_synth.info['chs'][2]['loc'] = np.array([-7.0025e-02, -3.6650e-03,  5.4835e-02, -7.4690e-02,  4.3000e-03, 4.5310e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][3]['loc'] = np.array([-7.0025e-02, -3.6650e-03,  5.4835e-02, -7.4690e-02,  4.3000e-03, 4.5310e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][4]['loc'] = np.array([-7.0885e-02, -2.0680e-02,  5.6790e-02, -7.6410e-02, -2.9730e-02, 4.9220e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][5]['loc'] = np.array([-7.0885e-02, -2.0680e-02,  5.6790e-02, -7.6410e-02, -2.9730e-02, 4.9220e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][6]['loc'] = np.array([-5.9145e-02, -2.0270e-02,  7.2330e-02, -5.2930e-02, -2.8910e-02, 8.0300e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synth.info['chs'][7]['loc'] = np.array([-5.9145e-02, -2.0270e-02,  7.2330e-02, -5.2930e-02, -2.8910e-02, 8.0300e-02, -6.5360e-02, -1.1630e-02,  6.4360e-02,  7.6000e+02, np.nan, np.nan])
    raw_synths[i] = raw_synth
# Show first raw object as example
raw_synths[0]

# Look at segments
events_synths, event_dict_synths = dict(), dict()
for i in range(n_set):
    events_synth, event_dict_synth = mne.events_from_annotations(raw_synths[i])
    events_synths[i] = events_synth
    event_dict_synths[i] = event_dict_synth
# Show events first raw object as example
plt.rcParams["figure.figsize"] = (10, 6) # (w, h)
fig = mne.viz.plot_events(events_synths[0], event_id=event_dict_synths[0], sfreq=raw_synths[0].info['sfreq'])
if export:
    fig.savefig('events_synth.png')

Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']

# Plot channels in 2D using build-in function of mne_nirs
sns.reset_defaults() # Turn of sns to be able to also plot contour of head
# dots = Center of channels
%matplotlib inline
fig = raw_synths[0].plot_sensors();
if export:
    fig.savefig('montage_synth.png')

sns.set_theme()

# Set constants 
sigma = 0.33 # Spatial correlation between channels
ar_order= 30 # Temporal correlation

# Define functions
def randAR(P):
    # random Pth order AR coef    
    a = np.flipud(np.cumsum(np.random.rand(P)))
    a = a / np.sum(a) * 0.99
    return a

def simARNoise(nchan, t, P=30, sigma=0.33):

    nsignal = nchan*2
    
    # noise mean and spatial covariance
    mu = np.zeros(nsignal)
    S = toeplitz([1] + [sigma] * (nsignal - 1))
    
    e = multivariate_normal.rvs(mean=mu, cov=S, size=len(t))

    # add temporal covariance
    for i in range(e.shape[1]):
        a = randAR(P)
        e[:, i] = lfilter([1], np.hstack([1, -a]), e[:, i])
    
    # output
    data = e * 5e-7 # How strong is the noise?
    data[:,1::2] = data[:,1::2]/2 # HbR values lower than HbO
    
    return data.T

# Set colored noise as data of raw object
for i in range(n_set):
    raw_synths[i]._data = simARNoise(nchan, t, ar_order, sigma)

# Plot data
fig = raw_synths[0].plot(duration=totLength)
if export:
    fig.savefig('synth_colored_noise.png')

# Set constants
# Assume: heart rate ~1.3 Hz, respiration ~0.3 Hz, Meyer waves ~0.1 Hz
cardiacHz = 1.318
respHz = 0.3
mayerHz = 0.1

cardiacHz_sigma = 0.242
respHz_sigma = 0.1
mayerHz_sigma = 0.02

# Specify amplitudes
cardiac_amp = 0.4
resp_amp = 0.25
mayer_amp = 0.75

def simPhysioNoise_variableHz(raw_input, time, cardiac_amp=0.25, resp_amp=0.25, mayer_amp=0.25, cardiacHz=1.318, cardiacHz_sigma = 0.242, respHz=0.3, respHz_sigma = 0.1, mayerHz=0.1, mayerHz_sigma=0.2, plot = False, tmin = 0, tmax = 10, ppf = 6):
    raw = raw_input.copy()
    
    if (raw_synth.ch_names[0][-3:] != 'hbo') or (raw_synth.ch_names[1][-3:] != 'hbr'):
        raw = beer_lambert_law(raw, ppf)
    
    nsignal, nsamp = raw.get_data().shape
    nchan = int(nsignal/2)
    
    # Generate cardiac oscillations
    cardiac_freq = np.random.normal(loc=cardiacHz, scale=cardiacHz_sigma)
    cardiac_phase = np.cumsum(0.1 * np.random.randn(nsamp) * 2 * np.pi / raw.info['sfreq'])
    cardiac_data = np.sin(2 * np.pi * cardiac_freq * time + cardiac_phase)

    # Generate respiratory oscillations
    resp_freq = np.random.normal(loc=respHz, scale=respHz_sigma)
    resp_phase = np.cumsum(0.1 * np.random.randn(nsamp) * 2 * np.pi / raw.info['sfreq'])
    resp_data = np.sin(2 * np.pi * resp_freq * time + resp_phase)

    # Generate Mayer waves
    mayer_freq = np.random.normal(loc=mayerHz, scale=mayerHz_sigma)
    mayer_phase = np.cumsum(0.1 * np.random.randn(nsamp) * 2 * np.pi / raw.info['sfreq'])
    mayer_data = np.sin(2 * np.pi * mayer_freq * time + mayer_phase)

    # Extract data from raw object
    Y = raw.get_data()

    
    # Add physiological noise
    if plot:
        fig, axes = plt.subplots(nchan, 2, figsize=(14, 5*nchan), constrained_layout=True)
        fig.suptitle('Physiological noise introduction to synthetic data', fontweight = 'bold')
        
        for k, (chan, ax) in enumerate(zip(raw.ch_names, axes.flat)):
            color1 = 'm' if k%2 == 1 else 'orange'
            color2 = 'blue' if k%2 == 1 else 'red'
            ax.plot(t, 1e6*Y[k], '--', color = color1, label = 'original')
            sigma = np.std(Y[k], axis=0)
            Y[k]  += cardiac_amp * sigma * cardiac_data + resp_amp * sigma * resp_data + mayer_amp * sigma * mayer_data
            ax.plot(t, 1e6*Y[k], '-', color = color2, label = 'with physiological noise')
            ax.set_xlabel('time (s)')
            ax.set_ylabel('Concentration (microM)')
            ax.set_xlim(tmin, tmax)
            ax.set_title('Channel ' + chan)
            ax.legend()
    else:
        for k in range(nsignal):
            sigma = np.std(Y[k], axis=0)
            Y[k]  += cardiac_amp * sigma * cardiac_data + resp_amp * sigma * resp_data + mayer_amp * sigma * mayer_data
    
    # Load data back in raw object
    raw._data = Y

    return raw

# Add physiological noise
raw_synth_physios = dict()
raw_synth_physios[0] = simPhysioNoise_variableHz(raw_synths[0], t, cardiac_amp, resp_amp, mayer_amp, cardiacHz, cardiacHz_sigma, respHz, respHz_sigma, mayerHz, mayerHz_sigma, plot = True, tmin = 30, tmax = 90, ppf = 6)
# Do not plot for all sets
for i in range(1, n_set):
    raw_synth_physios[i] = simPhysioNoise_variableHz(raw_synths[i], t, cardiac_amp, resp_amp, mayer_amp, cardiacHz, cardiacHz_sigma, respHz, respHz_sigma, mayerHz, mayerHz_sigma, plot = False, tmin = 30, tmax = 90, ppf = 6)

# Plot data
%matplotlib inline
fig = raw_synth_physios[0].copy().plot(duration=totLength)
if export:
    fig.savefig('synth_physio_noise.png')

# Generate HRFs
design_matrix = make_first_level_design_matrix(raw_synth_physios[0], hrf_model='glover', drift_order=0, stim_dur=stimDur)

# Overwrite the first NIRS channel with the expected response.
# Rescale to be in expected units of uM.
hrf = raw_synth_physios[0].copy().pick(picks=[0])
hrf._data[0] = 1e-6 * (design_matrix['Task']).T

hrf_hbo = hrf.get_data()[0]
hrf_hbr = -hrf.get_data()[0]/2

# Plot HRFs
fig = plt.figure(figsize=(12,2))
plt.plot(t, 1e6*hrf_hbo, color = 'r', label = 'HbO')
plt.plot(t, 1e6*hrf_hbr, color = 'b', label = 'HbR')
plt.xlabel('time (s)')
plt.ylabel('Concentration (microM)')
plt.legend()
plt.title('HRF of synthetic data', fontweight = 'bold')
plt.plot()
if export:
    fig.savefig('synth_HRF.png')

# Add HRF
hrf = np.zeros_like(raw_synth_physios[0].get_data())
hrf[::2,:] = hrf_hbo
hrf[1::2,:] = hrf_hbr
for i in range(n_set):
    raw_synth_physios[i]._data += hrf

fig = raw_synth_physios[0].plot(duration=totLength)
if export:
    fig.savefig('synth_HRF_added.png')

def inverse_beer_lambert(raw_input, ppf = 6):
    """
    Converts HbO and HbR data to optical density data using the inverse Beer-Lambert Law.

    Parameters:
    raw
    ppf (float or ndarray): Partial pathlength factor(s).

    Returns:
    opt_density (ndarray): Optical density data.
    """
    raw = raw_input.copy()
    data = raw.get_data()
    
    picks = np.arange(len(raw.ch_names))
    distances = source_detector_distances(raw.info, picks="all")
    
    freqs = np.array([raw.info["chs"][pick]["loc"][9] for pick in picks], float)
    abs_coef = _load_absorption(freqs)
    
    ppf = np.array([ppf, ppf])

    rename = dict()
    # Calculate optical density  
    for ii, jj in zip(picks[::2], picks[1::2]):
        EL = abs_coef * distances[ii] * ppf
        iEL = np.linalg.pinv(EL)

        raw._data[[ii, jj]] = np.linalg.pinv(iEL) @ (raw._data[[ii, jj]] * 1e3)

        # Update channel information
        coil_dict = dict(hbo=FIFF.FIFFV_COIL_FNIRS_HBO, hbr=FIFF.FIFFV_COIL_FNIRS_HBR)
        for ki, kind in zip((ii, jj), ("760", "850")):
            ch = raw.info["chs"][ki]
            ch.update(coil_type=FIFF.FIFFV_COIL_FNIRS_OD, unit=FIFF.FIFF_UNIT_V)
            new_name = f'{ch["ch_name"].split(" ")[0]} {kind}'
            rename[ch["ch_name"]] = new_name
    raw.rename_channels(rename)


    return raw

raw_synth_physio_ods = dict()
for i in range(n_set):
    raw_synth_physio_ods[i] = inverse_beer_lambert(raw_synth_physios[i], ppf = ppf)

%matplotlib inline
# Where do you want the MAs to be? No artefacts during baseline (Participant is asked to sit still)
motion_mask = np.ones(tot_n_samples)
s_start = int(t_begin*sfreq)
s_end = int((t_begin + baseDur)*sfreq)
indices = np.arange(s_start, s_end)
motion_mask[indices] = 0
plt.plot(t, hrf_hbo, label =  'HRF hbo')
plt.plot(t, 1e-6*motion_mask, label = 'mask')
plt.legend()
plt.xlabel('time (s)')
plt.ylabel('Concentration (M)')
plt.show()

# Set constants

# Frequency artefacts
spikes_per_minute = 3
shifts_per_minute = 1

# Amplitude artefacts (V)
amp_low_spike = 0.1
amp_high_spike = 0.5
amp_low_shift = 0.1
amp_high_shift = 0.5

# Define where the MAs are present
motion_mask = motion_mask

def sim_motion_artifact(raw_input, time, spikes_per_minute=3, shifts_per_minute=0.5, amp_low_spike = 0.1, amp_high_spike = 2, amp_low_shift = 0.1, amp_high_shift = 2, motion_mask=None):
    
    raw = raw_input.copy()
    data = raw.get_data()

    events_, event_dict_ = mne.events_from_annotations(raw)
    
    if motion_mask is None:
        motion_mask = np.ones_like(t, dtype=bool)

    num_spikes = int(round(spikes_per_minute * (time[-1] - time[0]) / 60))
    num_shifts = int(round(shifts_per_minute * (time[-1] - time[0]) / 60))

    valid_indices = np.where(motion_mask)[0]
    nsamp = len(valid_indices)
    inds = np.arange(0, nsamp)

    # MNE-NIRS can not handle two events at the same time
    # Convert event times to indices
    event_indices = np.intersect1d(valid_indices, events_[:, 0], assume_unique=True)
    
    # Exclude event indices from valid indices
    available_indices = np.setdiff1d(valid_indices, event_indices, assume_unique=True)
    
    # Select spike indices
    spike_inds = np.random.choice(available_indices, size=num_spikes, replace=False)
    
    # Remove spike indices from available indices for shift selection
    available_indices = np.setdiff1d(available_indices, spike_inds, assume_unique=True)
    
    # Select shift indices
    shift_inds = np.random.choice(available_indices, size=num_shifts, replace=False)

    spike_amp_Z = np.random.uniform(low=amp_low_spike, high=amp_high_spike, size=(len(data), num_spikes)) # Different amplitude for each channel
    shift_amp_Z = np.random.uniform(low=amp_low_shift, high=amp_high_shift, size=(len(data), num_spikes)) # Different amplitude for each channel

    mu = np.mean(data, axis = 1).reshape(-1, 1)
    # std = np.std(data, axis = 1).reshape(-1, 1)
    
    for i in range(num_spikes):
        width = 0.9 * np.random.rand() + 0.1  # Spike duration of 0.1-1 seconds

        t_peak = t[spike_inds[i]]
        t_start = t_peak - width / 2
        t_end = t_peak + width / 2
        t_inds = np.where((time > t_start) & (time < t_end))[0]
        amp = spike_amp_Z[:,i] * np.random.choice(np.array([-1,1])) # Positive or negative shift
        # amp = amp + (2 * np.random.rand(len(t_inds)) - 1) * 0.5 * amp  # Add temporal jitter

        spike_data = np.ones((len(data),len(t_inds))) * amp.reshape(-1, 1)
        t_inds = np.arange(t_inds[0], min(t_inds[0] + spike_data.shape[1], data.shape[1]))
        data[:,t_inds] = data[:,t_inds] + spike_data[:,:len(t_inds)]

        raw.annotations.append(t_start, width, 'Spike') # annotation during spike

    for i in range(num_shifts):
        shift_amt = shift_amp_Z[:,i] * np.random.choice(np.array([-1,1])) # Positive or negative shift
        data[:,shift_inds[i]:] = data[:,shift_inds[i]:] + shift_amt.reshape(-1, 1)

        raw.annotations.append(time[shift_inds[i]], 2, 'Shift') # 1 s annotation at baseline shift

    # Restore original mean intensity
    data = (data - np.mean(data, axis = 1).reshape(-1, 1)) + mu
    
    # Load data back in raw object
    raw._data = data
    return raw

raw_synth_physio_MAs = dict()
for i in range(n_set):
    raw_synth_physio_MAs[i] = sim_motion_artifact(raw_synth_physio_ods[i], t, spikes_per_minute, shifts_per_minute, amp_low_spike, amp_high_spike, amp_low_shift, amp_high_shift, motion_mask)

Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Task']

# Test if there are no duplicates in event data

for i in range(n_set):   
    events_, event_dict_ = mne.events_from_annotations(raw_synth_physio_MAs[i])
    while len(find_duplicates_with_set(events_[:,0])) > 0:
        print('Set: ' + str(i))
        print('Duplicates are at timepoint: ' + str(np.array(find_duplicates_with_set(events_[:,0]))/sfreq))
        raw_synth_physio_MAs[i] = sim_motion_artifact(raw_synth_physio_ods[i], t, spikes_per_minute, shifts_per_minute, amp_low_spike, amp_high_spike, amp_low_shift, amp_high_shift, motion_mask)
        events_, event_dict_ = mne.events_from_annotations(raw_synth_physio_MAs[i])

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Set: 5
Duplicates are at timepoint: [364.96598639]
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Set: 6
Duplicates are at timepoint: [295.06802721]
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Set: 7
Duplicates are at timepoint: [434.01360544]
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Set: 7
Duplicates are at timepoint: [390.47619048]
Used Annotations descriptions: ['Baseline', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']

# Test if there are no duplicates in event data
for i in range(n_set):   
    events_, event_dict_ = mne.events_from_annotations(raw_synth_physio_MAs[i])
    if len(find_duplicates_with_set(events_[:,0])) > 0:
        raise Exception('Set ' + str(i) + ' still contains uplicates at timepoint(s): ' + str(np.array(find_duplicates_with_set(events_[:,0]))/sfreq))

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']

# Use to rerun a set if there are duplicates
raw_synth_physio_MAs[3] = sim_motion_artifact(raw_synth_physio_ods[5], t, spikes_per_minute, shifts_per_minute, amp_low_spike, amp_high_spike, amp_low_shift, amp_high_shift, motion_mask)

Used Annotations descriptions: ['Baseline', 'Task']

raw_synth_od_unfiltereds = dict()
raw_synth_od_cleans = dict()
for i in range(n_set):
    raw_synth_od_unfiltereds[i] = raw_synth_physio_MAs[i].copy()
    raw_synth_od_cleans[i] = raw_synth_physio_ods[i].copy()

%matplotlib inline
sns.set()
# Look at segments
for i in range(n_set):
    events_synths[i], event_dict_synths[i] = mne.events_from_annotations(raw_synth_od_unfiltereds[i])
# Plot for first element
plt.rcParams["figure.figsize"] = (10, 6) # (w, h)
fig = mne.viz.plot_events(events_synths[0], event_id=event_dict_synths[0], sfreq=raw_synth_od_unfiltereds[0].info['sfreq'])
if export:
    fig.savefig('events_synth_MA.png')

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']

if plot:
    %matplotlib qt
    raw_synth_od_unfiltereds[0].plot(duration=120, show_scrollbars = True, clipping = None);

# Calculate SCI with build-in function of mne_nirs
sci_synth_unfiltered = scalp_coupling_index(raw_synth_od_unfiltereds[0], l_freq=0.7, h_freq=1.5, l_trans_bandwidth=0.3, h_trans_bandwidth=0.3, verbose=False)
# SCI bandpasses the signal with a FIR filter: l_freq=0.7, h_freq=1.5, l_trans_bandwidth=0.3, h_trans_bandwidth=0.3

# Copy raw object to apply SCI 
raw_synth_od_unfiltered_SCI = raw_synth_od_unfiltereds[0].copy()
# Mark channels with SCI < 0.75 as BAD
raw_synth_od_unfiltered_SCI.info['bads'] = list(compress(raw_synth_od_unfiltereds[0].ch_names, sci_synth_unfiltered < 0.75)) # 0.75 RECOMMENDED BY Pollonini et al. (2013)
print('The ' + str(len(raw_synth_od_unfiltered_SCI.info['bads'])) + ' bad channels are: ' + str(raw_synth_od_unfiltered_SCI.info['bads']))

The 0 bad channels are: []

# Calculate windowed SCI with windows of 30 s
_, scores, times = scalp_coupling_index_windowed(raw_synth_od_unfiltered_SCI, time_window=30)

%matplotlib inline
# Plot
plot_timechannel_quality_metric(raw_synth_od_unfiltered_SCI, scores, times, threshold=0.75, title="Scalp Coupling Index Quality Evaluation");

# View instances where a subset of channels may be contaminated by artifacts for a short duration of the recording
raw_synth_od_unfiltered_SCI_PSP, scores, times = peak_power(raw_synth_od_unfiltered_SCI, time_window=10)

# Plot
%matplotlib inline
fig = plot_timechannel_quality_metric(raw_synth_od_unfiltered_SCI_PSP, scores, times, threshold=0.1, title="Peak Power Quality Evaluation") #see PHOEBE
if export:
    fig.savefig('synth_PSP.png')

raw_synth_haemo_unfiltereds = dict()
raw_synth_haemo_cleans = dict()
for i in range(n_set):
    raw_synth_haemo_unfiltereds[i] = beer_lambert_law(raw_synth_od_unfiltereds[i], ppf=ppf)
    raw_synth_haemo_cleans[i] = beer_lambert_law(raw_synth_od_cleans[i], ppf=ppf)

# Plot concentrations
if plot:
    %matplotlib qt
    raw_synth_haemo_cleans[0].plot(duration=60, show_scrollbars = True, clipping = None);

# Plot concentrations
if plot:
    %matplotlib qt
    raw_synth_haemo_unfiltereds[0].plot(duration=60, show_scrollbars = True, clipping = None);

def get_DeltaSNR(raw_clean, raw_filtered, raw_unfiltered, plot = False, export = False, filename = None):
    """
    Input:
    - clean raw OD object
    - filtered raw OD object
    - unfiltered raw OD object 
    Output:
    - DeltaSNR value per channel
    """
    
    x = raw_clean.copy().get_data()
    x_hat =  raw_filtered.copy().get_data()
    z = raw_unfiltered.copy().get_data()
    
    SNR_e = 10 * np.log10(np.var(x, axis = 1)/np.var(x_hat-x, axis = 1))
    SNR_i = 10 * np.log10(np.var(x, axis = 1)/np.var(z-x, axis = 1))
    DeltaSNR = SNR_e - SNR_i

    if plot:
        %matplotlib inline
        y_pos = range(len(raw_clean.ch_names))
        fig = plt.figure(figsize=(12,6))
        plt.bar(y_pos, DeltaSNR)
        # Rotation of the bars names
        plt.xticks(y_pos, raw_clean.ch_names, rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('SNR value (dB)')
        plt.title('SNR per channel')
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('DeltaSNR/' + filename + '.png')

    return DeltaSNR

def get_PRD(raw_clean, raw_filtered, raw_unfiltered, plot = False, export = False, filename = None):
    """
    Input:
    - clean raw OD object
    - filtered raw OD object
    - unfiltered raw OD object 
    Output:
    - PRD value per channel
    """
    
    x = raw_clean.copy().get_data()
    x_hat =  raw_filtered.copy().get_data()
    z = raw_unfiltered.copy().get_data()

    PRD = 100*np.sqrt(np.sum((x_hat - x)**2, axis = 1)/np.sum(x**2, axis = 1))

    if plot:
        %matplotlib inline
        y_pos = range(len(raw_clean.ch_names))
        fig = plt.figure(figsize=(12,6))
        plt.bar(y_pos, PRD)
        # Rotation of the bars names
        plt.xticks(y_pos, raw_clean.ch_names, rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('PRD value (%)')
        plt.title('PRD per channel')
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('PRD/' + filename + '.png')
            
    return PRD

def get_DeltaCC(raw_clean, raw_filtered, raw_unfiltered, plot = False, export = False, filename = None):
    """
    Input:
    - clean raw OD object
    - filtered raw OD object
    - unfiltered raw OD object 
    Output:
    - DeltaCC value per channel
    """

    x = raw_clean.copy().get_data()
    x_hat =  raw_filtered.copy().get_data()
    z = raw_unfiltered.copy().get_data()

    DeltaCC = np.zeros(len(x))
    for chan in range(len(x)):
        DeltaCC[chan] = pearsonr(x_hat[chan], x[chan])[0] - pearsonr(z[chan], x[chan])[0] # Use pearson r correlation

    if plot:
        %matplotlib inline
        y_pos = range(len(raw_clean.ch_names))
        fig = plt.figure(figsize=(12,6))
        plt.bar(y_pos, DeltaCC)
        # Rotation of the bars names
        plt.xticks(y_pos, raw_clean.ch_names, rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('DeltaCC value')
        plt.title('DeltaCC per channel')
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('DeltaCC/' + filename + '.png')
            
    return DeltaCC

def get_CNR_Cooper(raw_haemo_filtered, raw_haemo_unfiltered, reject_criteria = dict(hbo=80e-6), tmin = -5, tmax = 15, t_peak = 7, real_HRF= None, plot = False, export = False, filename = None):
    """
    Based on HRF!
    Input:
    - filtered raw haemo object
    - unfiltered raw haemo object 
    Output:
    - CNR value per channel 
    """

    raw_haemo_filtered_, raw_haemo_unfiltered_ = raw_haemo_filtered.copy(), raw_haemo_unfiltered.copy()
    events_, event_dict_ = mne.events_from_annotations(raw_haemo_filtered_)
    
    # Create epochs
    epochs_filtered = mne.Epochs(raw_haemo_filtered_, events_, event_dict_, tmin, tmax, preload=True, reject=reject_criteria, reject_by_annotation=True, baseline=(None, 0))
    epochs_unfiltered = mne.Epochs(raw_haemo_unfiltered_, events_, event_dict_, tmin, tmax, preload=True, reject=reject_criteria, reject_by_annotation=True, baseline=(None, 0))
    
    HRF_hbo_filtered = epochs_filtered["Task"].average(picks="hbo").get_data()
    HRF_hbo_unfiltered = epochs_unfiltered["Task"].average(picks="hbo").get_data()

    HRF_hbr_filtered = epochs_filtered["Task"].average(picks="hbr").get_data()
    HRF_hbr_unfiltered = epochs_unfiltered["Task"].average(picks="hbr").get_data()
    
    if plot:   
        t = np.arange(HRF_hbo_filtered.shape[1])/raw_haemo_filtered_.info['sfreq'] + tmin
        for chan in range(len(HRF_hbo_filtered)):
            plt.vlines([t_peak-1, t_peak+1], linestyle = '-.', ymin = np.min(1e6*HRF_hbr_unfiltered[chan])-1, ymax = np.max(1e6*HRF_hbo_unfiltered[chan])+1, label = '2 seconds interval around peak HRF', color = 'k')
            plt.vlines([-5, 0], linestyle = '-.', ymin = np.min(1e6*HRF_hbr_unfiltered[chan])-1, ymax = np.max(1e6*HRF_hbo_unfiltered[chan])+1, label = '5 seconds interval before onset stimulus', color = 'orange')
            plt.plot(t, 1e6*HRF_hbo_filtered[chan], label = 'HRF hbo filtered', color = 'r')
            plt.plot(t, 1e6*HRF_hbr_filtered[chan], label = 'HRF hbr filtered', color = 'b')
            plt.plot(t, 1e6*HRF_hbo_unfiltered[chan], '--', label = 'HRF hbo unfiltered', color = 'r')
            plt.plot(t, 1e6*HRF_hbr_unfiltered[chan], '--', label = 'HRF hbr unfiltered', color = 'b')
            if not real_HRF.all() == None:
                plt.plot(t, 1e6*real_HRF, '-.', label = 'real HRF hbo', color = 'r')
                plt.plot(t, -5e5*real_HRF, '-.', label = 'real HRF hbr', color = 'b')
            plt.xlabel('time (s)')
            plt.ylabel('Concentration (µM)')
            plt.title('Channel ' + str(chan+1))
            plt.legend()
            plt.show()
            if export:
                if filename == None:
                    raise ValeError('Add filename')
                plt.savefig('CNR_Cooper_functioning_' + filename + '.png')

    mean_HRF_peak_filtered = np.mean(epochs_filtered["Task"].crop(t_peak-1,t_peak+1).average(picks="hbo").get_data(), axis = 1)
    std_baseline_filtered = np.std(epochs_filtered["Task"].crop(-5,0).average(picks="hbo").get_data(), axis = 1)

    CNR_filtered = mean_HRF_peak_filtered/std_baseline_filtered

    mean_HRF_peak_unfiltered = np.mean(epochs_unfiltered["Task"].crop(t_peak-1,t_peak+1).average(picks="hbo").get_data(), axis = 1)
    std_baseline_unfiltered = np.std(epochs_unfiltered["Task"].crop(-5,0).average(picks="hbo").get_data(), axis = 1)

    CNR_unfiltered = mean_HRF_peak_unfiltered/std_baseline_unfiltered    

    if plot:
        %matplotlib inline
        x_pos = np.arange(len(CNR_filtered))
        width = 0.2
        fig = plt.figure(figsize=(12,6))
        plt.bar(x_pos - width/2, CNR_filtered, width=width, label='Filtered')
        plt.bar(x_pos + width/2, CNR_unfiltered, width=width, label='Unfiltered')
        # Rotation of the bars names
        plt.xticks(x_pos, raw_haemo_filtered_.ch_names[::2], rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('CNR (Cooper) value')
        plt.title('CNR (Cooper) per channel')
        plt.legend()
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('CNR_C/' + filename + '.png')
            
    return CNR_filtered

def get_CNR_Zhang(raw_haemo_filtered, raw_haemo_unfiltered, reject_criteria = dict(hbo=80e-6), epsilon = 1e-7, plot = False, export = False, filename = None):
    """
    Input:
    - filtered raw haemo object
    - unfiltered raw haemo object
    Output:
    - CNR value per channel
    """

    raw_haemo_filtered_, raw_haemo_unfiltered_ = raw_haemo_filtered.copy(), raw_haemo_unfiltered.copy() 
    events_, event_dict_ = mne.events_from_annotations(raw_haemo_filtered_)
    
    tmin = -15
    tmax = 20

    # Create epochs
    epochs_filtered = mne.Epochs(raw_haemo_filtered_, events_, event_dict_, tmin, tmax, preload=True, reject=reject_criteria, reject_by_annotation=True, baseline=(None, 0))
    epochs_unfiltered = mne.Epochs(raw_haemo_unfiltered_, events_, event_dict_, tmin, tmax, preload=True, reject=reject_criteria, reject_by_annotation=True, baseline=(None, 0))

    # Take pre and dur segment
    HRF_hbo_pre_filtered = epochs_filtered["Task"].average(picks="hbo").copy().crop(-10,0).get_data()
    HRF_hbo_dur_filtered = epochs_filtered["Task"].average(picks="hbr").copy().crop(5,15).get_data()

    HRF_hbo_pre_unfiltered = epochs_unfiltered["Task"].average(picks="hbo").copy().crop(-10,0).get_data()
    HRF_hbo_dur_unfiltered = epochs_unfiltered["Task"].average(picks="hbr").copy().crop(5,15).get_data()

    CNR_filtered = (np.mean(HRF_hbo_dur_filtered, axis = 1) - np.mean(HRF_hbo_pre_filtered, axis = 1))/np.sqrt(np.var(HRF_hbo_dur_filtered, axis = 1) + np.var(HRF_hbo_pre_filtered, axis = 1) + epsilon)
    CNR_unfiltered = (np.mean(HRF_hbo_dur_unfiltered, axis = 1) - np.mean(HRF_hbo_pre_unfiltered, axis = 1))/np.sqrt(np.var(HRF_hbo_dur_unfiltered, axis = 1) + np.var(HRF_hbo_pre_unfiltered, axis = 1) + epsilon)
    
    if plot:
        %matplotlib inline
        x_pos = np.arange(len(CNR_filtered))
        width = 0.2
        fig = plt.figure(figsize=(12,6))
        plt.bar(x_pos - width/2, CNR_filtered, width=width, label='Filtered')
        plt.bar(x_pos + width/2, CNR_unfiltered, width=width, label='Unfiltered')
        # Rotation of the bars names
        plt.xticks(x_pos, raw_haemo_filtered_.ch_names[::2], rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('CNR (Zhang) value')
        plt.title('CNR (Zhang) per channel')
        plt.legend()
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('CNR_Z/' + filename + '.png')
            
    return CNR_filtered

def get_APA(raw_filtered_, raw_unfiltered_, Am, l_freq, filter_length='auto', l_trans_bandwidth=0.3, method='fir', iir_params=None, phase='zero', fir_window='hamming', fir_design='firwin', plot = False, export = False, filename = None):
    """
    Input:
    - filtered raw OD object
    - filtered raw OD object
    - set of indices for time segments where MAs occur. For n segments: Am = [(start_1, stop_1), ..., (start_n, stop_n)]
    - l_freq: For FIR filters, the lower pass-band edge; for IIR filters, the lower cutoff frequency.
    - l_trans_bandwidth: transition band at the low cut-off frequency in Hz 
    - ... Other optonal parameters to specify filter, see mne.io.raw.filter
    Output:
    - APA value per channel
    """

    raw_filtered, raw_unfiltered = raw_filtered_.copy(), raw_unfiltered_.copy()
    
    if plot:
        filter_params = create_filter(raw_filtered.get_data(), raw_filtered.info['sfreq'], l_freq=l_freq, h_freq = None, l_trans_bandwidth=l_trans_bandwidth)
        mne.viz.plot_filter(filter_params, raw_filtered.info['sfreq'], fscale='log', gain=False, plot='magnitude')
    
    x_hat_HP = raw_filtered.filter(l_freq, None, filter_length=filter_length, l_trans_bandwidth=l_trans_bandwidth, method=method, iir_params=iir_params, phase=phase, fir_window=fir_window, fir_design=fir_design)
    z_HP = raw_unfiltered.filter(l_freq, None, filter_length=filter_length, l_trans_bandwidth=l_trans_bandwidth, method=method, iir_params=iir_params, phase=phase, fir_window=fir_window, fir_design=fir_design)

    z_HP_sq = z_HP.get_data()**2
    x_hat_HP_sq = x_hat_HP.get_data()**2

    APA = np.zeros(len(raw_filtered.ch_names))
    for segm in Am:
        top = np.mean(z_HP_sq[:,segm[0]:segm[1]], axis = 1)
        bottom = np.mean(x_hat_HP_sq[:,segm[0]:segm[1]], axis = 1)
        apa = 10 * np.log10(top/bottom)
        APA += apa # Sum score of each artefact to get 1 APA score per channel

    if plot:
        %matplotlib inline
        x_pos = range(len(raw_filtered.ch_names))
        fig = plt.figure(figsize=(12,6))
        plt.bar(x_pos, APA)
        # Rotation of the bars names
        plt.xticks(x_pos, raw_filtered.ch_names, rotation=90)
        plt.xlabel('Channel name')
        plt.ylabel('APA value of artefacts summed per channel (dB)')
        plt.title('APA per channel')
        plt.show()
        if export:
            if filename == None:
                raise ValeError('Add filename')
            fig.savefig('APA/' + filename + '.png')
    
    return APA

# Define help functions
def f1(x):
    if x >= 0:
        result = 1
    else: 
        result = 0
    return result

def f2(x):
    if x > 0:
        result = 1
    else: 
        result = 0
    return result

def flag_ME(a_x, a_y, n, fs):
    DeltaT = 1/fs
    cond_1 = f1(np.abs(a_x[n]-a_x[n-1]) - 1.3*DeltaT)
    cond_2 = f1(np.abs(a_y[n]-a_y[n-1]) - 1.3*DeltaT)
    result = cond_1 or cond_2
    return result

def find_ME(a_x, a_y, fs, plot = False):
    ME_times = []
    for n in range(1, len(a_x)):
        score = flag_ME(a_x, a_y, n, fs)
        if score:
            ME_times.append(n/fs)
    ME_intervals = []
    for i in range(len(ME_times)):
        if i  == 0:
            ME_intervals.append(ME_times[i])
        elif (ME_times[i] - ME_times[i-1]) > 5: # Put shorter as MA are clode to each other, orgininal 20 s
            if ME_times[i-1] != ME_intervals[-1]:
                ME_intervals.append(ME_times[i-1])
            ME_intervals.append(ME_times[i])
    ME_intervals.append(ME_times[-1])

    ME_start = []
    for T_start in ME_intervals[::2]:
        ME_start.append(T_start)
    
    ME_duration = []
    for i in range(len(ME_intervals[::2])):
        T_m = ME_intervals[1::2][i] - ME_intervals[::2][i]
        if T_m >= 1: # (*)
            ME_duration.append(T_m)
        else:
            ME_start.pop(i)
            

    flagged = np.ones_like(a_x)*(-1)
    k = -1
    for i in range(len(ME_start)):
        s_start = round(ME_start[i]*fs)
        s_end = round((ME_start[i]+ME_duration[i])*fs) + 1
        k += 1
        flagged[s_start:s_end] = k
        
    if plot:
        t = np.arange(len(a_x))/fs
        ymin = np.min(np.concatenate([a_x, a_y]))
        ymax = np.max(np.concatenate([a_x, a_y]))
        plt.figure(figsize=(12,6))
        plt.plot(t, a_x, label = 'acc X')
        plt.plot(t, a_y, label = 'acc Y')
        plt.plot(ME_times, len(ME_times)*[ymax/4], 'x', color = 'y', label = 'ME')
        plt.vlines(ME_intervals, ymin, ymax, 'g', label = 'ME interval')
        plt.plot(t, flagged)
        #plt.vlines(ME_times, ymin, ymax)
        plt.legend()
        plt.xlabel('time (s)')
        plt.ylabel('g')
        plt.show()
    
    return flagged, ME_start

def avg_before(Z, T_start, fs):
    T_before = T_start - 1 # lower value than original code
    smin = round((T_before-3)*fs)
    if smin < 0:
        smin = 0
    smax = round(T_before*fs)
    if smax > len(Z[0]):
        smax = len(Z[0])
    return np.mean(Z[:,smin:smax], axis = -1)

def avg_after(Z, T_start, fs):
    T_after = T_start + 1
    smin = round(T_after*fs)
    if smin < 0:
        smin = 0
    smax = round((T_after+3)*fs)
    if smax > len(Z[0]):
        smax = len(Z[0])
    return np.mean(Z[:,smin:smax], axis = -1)

def std_before(Z, T_start, fs):
    T_before = T_start - 1 
    smin = round((T_before-3)*fs)
    if smin < 0:
        smin = 0
    smax = round(T_before*fs)
    if smax > len(Z[0]):
        smax = len(Z[0])
    return np.std(Z[:,smin:smax], axis = -1)
    

def flag_BS(Z, T_start, fs):
    return np.where((np.abs(avg_before(Z, T_start, fs) - avg_after(Z, T_start, fs)) - 2.6*std_before(Z, T_start, fs)) >= 0, 1, 0)

def N_ch(Z, T_start, fs):
    result = np.sum(flag_BS(Z, T_start, fs))
    return result

def N_wv(Z, T_start, fs):
    cond1 = flag_BS(Z[1::2], T_start, fs) > 0
    cond2 = flag_BS(Z[::2], T_start, fs) > 0
    result = np.sum(np.where(np.logical_and(cond1, cond2), 1, 0))
    return result
    
def flag_corr(Z, a_x, a_y, n, fs):
    result = 0
    flagged, ME_start = find_ME(a_x, a_y, fs, plot = False)
    if flagged[n] >= 0:
        T_start = ME_start[int(flagged[n])]
        cond1 = f1(N_ch(Z, T_start, fs) - 2)
        cond2 = f1(N_wv(Z, T_start, fs))
        if cond1 or cond2:
            result = 1           
    return result

# Define ABAMAR
def ABAMAR(Z, a_x, a_y, fs, plot = True):
    Z_new = Z.copy()
    flagged, ME_start = find_ME(a_x, a_y, fs)
    n_MA = -1
    for n in range(len(a_x)):   
        if flag_corr(Z, a_x, a_y, n, fs):
            T_start = ME_start[int(flagged[n])]
            Z_new[:,n] = avg_before(Z, T_start, fs)
            n_MA = int(flagged[n])
        elif n_MA >= 0:  
            Z_new[:,n] = avg_before(Z, T_start, fs)/avg_after(Z, T_start, fs)*Z[:,n]
        # the amplitude baseline preceding the event is imposed on all channels and wavelengths by multiplying 
        # amplitude values after Tm by A_before/A_after. Amplitude values between Tbefore and Tafter are set to 
        # A_before to remove any transient motion artifact
    # -->  A_before/A_after still to complete
    if plot:
        for i in range(len(Z)):
            t = np.arange(len(Z[i]))/fs
            plt.plot(t, Z[i], label = 'old')
            plt.plot(t, Z_new[i], '--', label = 'new')
            plt.legend()
            plt.xlabel('time (s)')
            plt.title('Channel ' + str(i))
            plt.show()
    return Z_new

#tmin, tmax = 105, 195
# s_start = round(tmin*raw_MA_od.info['sfreq'])
# s_end = round(tmax*raw_MA_od.info['sfreq']) + 1
# a_x, a_y = acc_df.to_numpy().T[:2, s_start:s_end]/9.81 # Convert to g instead of m/s^2
# a_x, a_y = acc_df.to_numpy().T[:2]/9.81
# raw_MA_od_ABAMAR = raw_od.copy()
# Z = raw_MA_od_ABAMAR.get_data()
# # Z = raw_MA_od.copy().crop(tmin, tmax).get_data()
# _, __ = find_ME(a_x, a_y, sfreq, plot = True)
# Z_new = ABAMAR(Z, a_x, a_y, sfreq, plot = True)
# raw_MA_od_ABAMAR._data = Z_new

# get_IMU_data

def ANC(raw_input, mu, n, acc_df):
    """
    - n: the amount of filter inputs (size of input vector)
    """
    raw = raw_input.copy()
    
    nirs_data = raw.get_data()
    acc_data = acc_df.to_numpy().T
    
    # # removing the mean and scaling to unit variance
    # scaler_nirs = StandardScaler().fit(nirs_data)
    # scaler_acc = StandardScaler().fit(acc_data)
    # nirs_data = scaler_nirs.transform(nirs_data)
    # acc_data = scaler_acc.transform(acc_data)

# raw_od_filtered_ANC = ANC(raw_od, mu, n, txt_filename)

# %matplotlib qt
# raw_od.plot(duration = 500)

# %matplotlib inline
# tmin, tmax = 400, 500
# noisy_signal = raw_od.copy().crop(tmin, tmax).get_data(picks = ['S3_D7 760'])[0]
# s_start = round(tmin*raw_od.info['sfreq'])
# s_end = round(tmax*raw_od.info['sfreq']) + 1
# acc_data = acc_df['accY'].to_numpy()[s_start:s_end]
# # scale:
# acc_data = (acc_data - np.min(acc_data))/(np.max(acc_data)-np.min(acc_data))
# noisy_signal = (noisy_signal - np.min(noisy_signal))/(np.max(noisy_signal)-np.min(noisy_signal))
# plt.figure(figsize=(12,6))
# plt.plot(noisy_signal, label ='fnirs')
# plt.plot(acc_data, label = 'acc')
# plt.legend()
# plt.show()

# len(x[0])

# # signals creation: u, v, d

# n = 5
# x = acc_df.to_numpy()
# d = raw_od.copy().get_data(picks = ['S3_D7 760'])[0]

# # filtering
# #x = pa.input_from_history(d, n)[:-1]
# #d = d[n:]
# f = pa.filters.FilterRLS(mu=0.99999, n=n)
# y, e, w = f.run(x, d)


# # show results
# plt.figure(figsize=(15,9))
# plt.subplot(211);plt.title("Adaptation");plt.xlabel("samples - k")
# plt.plot(d, "b", label="d - target", linestyle = 'dashed')
# plt.plot(y,"g", label="y - output")
# plt.legend()
# plt.subplot(212);plt.title("Filter error");plt.xlabel("samples - k")
# plt.plot(10*np.log10(e**2),"r", label="e - error [dB]");plt.legend()
# plt.tight_layout()
# plt.show()

# plt.plot(w)

# raw_od.plot(duration = 600)

# # Simulate data
# np.random.seed(0)  # For reproducibility
# n_samples = 500
# t = np.linspace(0, 10, n_samples)
# real_signal = np.sin(t)  # Example real signal (hemodynamic response)
# noise_signal = 0.5 * np.random.normal(size=n_samples)  # Example noise signal
# measured_signal = real_signal + noise_signal  # Combined measured signal
# accelerometer_signal = noise_signal + 0.1 * np.random.normal(size=n_samples)  # Simulated accelerometer signal (motion)

# # Data for RLS filter
# df = pd.DataFrame({
#     'measured_signal': measured_signal,
#     'accelerometer_signal': accelerometer_signal,
# })

# # Initialize RLS filter
# filter_order = 5  # Order of the filter
# lambda_factor = 0.99  # Forgetting factor for RLS

# # Apply the RLS adaptive filter
# rls_filter = pa.filters.FilterRLS(n=filter_order, mu=lambda_factor)
# df['v_hat'] = rls_filter.run(df['accelerometer_signal'], df['measured_signal'])
# df['recovered_signal'] = df['measured_signal'] - df['v_hat']

# # Plot the results
# plt.figure(figsize=(12, 8))

# # Measured vs Recovered signal
# plt.subplot(2, 1, 1)
# plt.plot(t, real_signal, label='Real Signal', linestyle='--')
# plt.plot(t, measured_signal, label='Measured Signal')
# plt.plot(t, df['recovered_signal'], label='Recovered Signal', linestyle='-.')
# plt.xlabel('Time')
# plt.ylabel('Signal')
# plt.title('Measured and Recovered Signal')
# plt.legend()

# # Noise estimation
# plt.subplot(2, 1, 2)
# plt.plot(t, noise_signal, label='True Noise Signal', linestyle='--')
# plt.plot(t, df['v_hat'], label='Estimated Noise (v_hat)')
# plt.xlabel('Time')
# plt.ylabel('Noise')
# plt.title('Noise Estimation')
# plt.legend()

# plt.tight_layout()
# plt.show()

# Method relies on differences between artifacts and fNIRS signal in terms of duration and amplitude and is specifically designed for spike artifacts.
# DOI: 10.1088/0967-3334/33/2/259

# Define wavelet function for a signle channel
def wavelet(data, datatype, fs, wavelet_type = 'db5', mode = 'symmetric', level = None, alpha = 0.1, plot = False, pick = None, export = False, filename = None):
    
    # Describe properties of a discrete wavelet
    wavelet = pywt.Wavelet(wavelet_type)
    
    # Calculate approximation and detail coefficients (cA (LF), cD (HF))
    coeff = pywt.wavedec(data, wavelet, mode=mode, level=level, axis=-1) 
    # Daubechies 5 (db5) wavelet, If level is None (default) then it will be calculated using the dwt_max_level function.
    # dwt_max_level stops decomposition when the signal becomes shorter than the FIR filter length for a given wavelet.
    # Structure of coeff: [cA_n, cD_n, cD_n-1, …, cD2, cD1]
    
    # Create empty new coefficient array and fill first element (Ac)
    coeff_new = []
    coeff_new.append(coeff[0])
    
    # Remove Ac element for wavelet filtering
    coeff.pop(0)
    n = len(coeff)

    # Create empty variance array and fill with median absolute deviation (MAD) (Hoaglin et al 1983)
    sigma = np.zeros(n)
    for i, coefficient in enumerate(coeff):
        sigma[i] = np.median(abs(coefficient))/0.6745 
    
    # Artifacts are treated as large outliers added to the desired coefficients∼ N(0,σ2).
    for i in range(n):
        prob = 2*(1 - norm.cdf(abs(coeff[i]/sigma[i])))
        coeff_new.append(np.where(prob > alpha, coeff[i], np.zeros_like(coeff[i])))
        
    # Perform inverse wavelet transform
    data_new = pywt.waverec(coeff_new, wavelet, mode=mode, axis=-1)

    if plot:
        # Plot original and new values
        if pick == None:
            raise ValueError('Please add picks')
        if datatype == 'hbo':
            type = ' hbo' if pick%2 == 0 else ' hbr'
            ylabel = 'HbO/HbR (M)'
        else:
            type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
            ylabel = 'OD (V)'
        %matplotlib inline
        time1 = np.arange(len(data))/fs
        time2 = np.arange(len(data_new))/fs
        plt.plot(time1, data, label = 'original')
        plt.plot(time2, data_new, label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            plt.savefig('WF/WF_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

    return data_new

# Iterate over all channels and create new raw object
def wavelet_filter(raw, plot = False, export = False, filename = None):
    
    raw = raw.copy()
    picks = np.sort(_validate_nirs_info(raw.info))
    n_samples = raw.get_data().shape[1]
    type = raw.ch_names[0][-3:]

    for pick in picks:
        raw._data[pick] = wavelet(raw._data[pick], type, raw.info["sfreq"], wavelet_type = 'db5', mode = 'symmetric', level = None, alpha = 0.1, plot = plot, pick = pick, export = export, filename = filename)[:n_samples]
        
    return raw

raw_synth_od_filtered_WFs = dict()
raw_synth_od_filtered_WFs[0] = wavelet_filter(raw_synth_od_unfiltereds[0], plot = plot, export = export, filename = 'synthetic')
for i in range(1, n_set):
    raw_synth_od_filtered_WFs[i] = wavelet_filter(raw_synth_od_unfiltereds[i])

# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 10, 40
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_WFs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('WF/WF_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 310, 340
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_WFs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq +tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('WF/WF_synth_detail2.png')

# Show example of plots for html page
plot = True

# SNR (OD data)
DeltaSNR_synth_od_WFs = dict()
DeltaSNR_synth_od_WFs[0] = get_DeltaSNR(raw_synth_od_cleans[0], raw_synth_od_filtered_WFs[0], raw_synth_od_unfiltereds[0], plot = plot, export = export, filename = 'WF')
for i in range(1, n_set):
    DeltaSNR_synth_od_WFs[i] = get_DeltaSNR(raw_synth_od_cleans[i], raw_synth_od_filtered_WFs[i], raw_synth_od_unfiltereds[i])

# PRD (OD data)
PRD_synth_od_WFs = dict()
PRD_synth_od_WFs[0] = get_PRD(raw_synth_od_cleans[0], raw_synth_od_filtered_WFs[0], raw_synth_od_unfiltereds[0], plot = plot, export = export, filename = 'WF')
for i in range(1,n_set):
    PRD_synth_od_WFs[i] = get_PRD(raw_synth_od_cleans[i], raw_synth_od_filtered_WFs[i], raw_synth_od_unfiltereds[i])

# CC (OD data)
DeltaCC_synth_od_WFs = dict()
DeltaCC_synth_od_WFs[0] = get_DeltaCC(raw_synth_od_cleans[0], raw_synth_od_filtered_WFs[0], raw_synth_od_unfiltereds[0], plot = plot, export = export, filename = 'WF')
for i in range(1,n_set):
    DeltaCC_synth_od_WFs[i] = get_DeltaCC(raw_synth_od_cleans[i], raw_synth_od_filtered_WFs[i], raw_synth_od_unfiltereds[i])

# CNR Cooper (Haemo data)

# Define rejection criteria
reject_criteria = dict(hbo=80e-6)
# Define start and end of epoch
tmin_CNR, tmax_CNR = -6, 30
t_peak = 10

# get real HRF for hbo (possible for synthetic data)
ind_stim = events_synths[0][np.where(events_synths[0][:,2] == 4)[0][5]][0]
ind_start = ind_stim + int(tmin_CNR*sfreq)
ind_stop = ind_stim + int(tmax_CNR*sfreq)
real_HRF = hrf_hbo[ind_start:ind_stop+1]

raw_synth_haemo_filtered_WFs = dict()
for i in range(n_set):
    raw_synth_haemo_filtered_WFs[i] = beer_lambert_law(raw_synth_od_filtered_WFs[i], ppf=ppf)
CNR_cooper_synth_haemo_WFs = dict()
CNR_cooper_synth_haemo_WFs[0] = get_CNR_Cooper(raw_synth_haemo_filtered_WFs[0], raw_synth_haemo_unfiltereds[0], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF, plot = plot, export = export, filename = 'WF')
for i in range(1, n_set):
    CNR_cooper_synth_haemo_WFs[i] = get_CNR_Cooper(raw_synth_haemo_filtered_WFs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped

# CNR Zhang (haemo data)
CNR_zhang_synth_haemo_WFs = dict()
CNR_zhang_synth_haemo_WFs[0] = get_CNR_Zhang(raw_synth_haemo_filtered_WFs[0], raw_synth_haemo_unfiltereds[0], reject_criteria = reject_criteria, plot = plot, export = export, filename = 'WF')
for i in range(1, n_set):
    CNR_zhang_synth_haemo_WFs[i] = get_CNR_Zhang(raw_synth_haemo_filtered_WFs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped

# APA (OD data)
l_freq = 1.5 # Hz
l_trans_bandwidth = 0.3 # Hz

ids = [event_dict_synths[0]['Spike'], event_dict_synths[0]['Shift']]

start_ind = events_synths[0][np.isin(events_synths[0][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
Am_synth = zip(start_ind, stop_ind)

APA_synth_od_WFs = dict()
APA_synth_od_WFs[0] = get_APA(raw_synth_od_filtered_WFs[0], raw_synth_od_unfiltereds[0], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth, plot = True)

for i in range(1, n_set):
    ids = [event_dict_synths[i]['Spike'], event_dict_synths[i]['Shift']]
    start_ind = events_synths[i][np.isin(events_synths[i][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    Am_synth = zip(start_ind, stop_ind)
    APA_synth_od_WFs[i] = get_APA(raw_synth_od_filtered_WFs[i], raw_synth_od_unfiltereds[i], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

plot = False

raw_od_filtered_WF = wavelet_filter(raw_od, plot = plot, export = export, filename = 'MA')

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_WF.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('WF/WF_MA_detail.png')

S2_D2 760

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_WF.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('WF/WF_MA_detail2.png')

S2_D2 760

plot = True

# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_WF = get_APA(raw_od_filtered_WF, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth, plot = True)

Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

plot = False

# Define function for a single channel
def wavelet_k(data, datatype, fs, wavelet_type = 'db5', mode = 'symmetric', level = None, k = 3.3, plot = False, pick = None, export = False, filename = None):
    
    # Describe properties of a discrete wavelet
    wavelet = pywt.Wavelet(wavelet_type)
    
    # Calculate approximation and detail coefficients (cA (LF), cD (HF))
    coeff = pywt.wavedec(data, wavelet, mode=mode, level=level, axis=-1) 
    # Daubechies 5 (db5) wavelet, If level is None (default) then it will be calculated using the dwt_max_level function.
    # dwt_max_level stops decomposition when the signal becomes shorter than the FIR filter length for a given wavelet.
    # Structure of coeff: [cA_n, cD_n, cD_n-1, …, cD2, cD1]

    coeff_new = []
    
    for coefficient in coeff:
        # Ignore zero values
        zero_free = coefficient[coefficient != 0]

        while kurtosis(zero_free) > k:
            coefficient[np.argmax(np.absolute(coefficient))] = 0
            zero_free = coefficient[coefficient != 0]
        coeff_new.append(coefficient)
        
    #Perform inverse wavelet transform
    data_new = pywt.waverec(coeff_new, wavelet, mode=mode, axis=-1)
    
    if plot:
        # Plot original and new values
        if pick == None:
            raise ValueError('Please add picks')
        if datatype == 'hbo':
            type = ' hbo' if pick%2 == 0 else ' hbr'
            ylabel = 'HbO/HbR (M)'
        else:
            type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
            ylabel = 'OD (V)'
        %matplotlib inline
        time1 = np.arange(len(data))/fs
        time2 = np.arange(len(data_new))/fs
        plt.plot(time1, data, label = 'original')
        plt.plot(time2, data_new, label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            plt.savefig('kbWF/kbWF_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

    return data_new

# Iterate over all channels and create new raw object
def wavelet_filter_k(raw, k = 3.3, plot = False, export = False, filename = None):
    
    raw = raw.copy()
    picks = np.sort(_validate_nirs_info(raw.info))
    n_samples = raw.get_data().shape[1]
    type = raw.ch_names[0][-3:]
    
    for pick in picks:
        raw._data[pick] = wavelet_k(raw._data[pick], type, raw.info['sfreq'], wavelet_type = 'db5', mode = 'symmetric', level = None, k = k, plot = plot, pick = pick, export = export, filename = filename)[:n_samples]
    
    return raw

raw_synth_od_filtered_kWFs = dict()
raw_synth_od_filtered_kWFs[0] = wavelet_filter_k(raw_synth_od_unfiltereds[0], k = 3.3, plot = plot, export = export, filename = 'synthetic')
for i in range(1, n_set):
    raw_synth_od_filtered_kWFs[i] = wavelet_filter_k(raw_synth_od_unfiltereds[i], k = 3.3)

# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 10, 40
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_kWFs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('kbWF/kbWF_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 310, 340
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_kWFs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq +tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('kbWF/kbWF_synth_detail2.png')

# SNR (OD data)
DeltaSNR_synth_od_kWFs = dict()
for i in range(n_set):
    DeltaSNR_synth_od_kWFs[i] = get_DeltaSNR(raw_synth_od_cleans[i], raw_synth_od_filtered_kWFs[i], raw_synth_od_unfiltereds[i])

# PRD (OD data)
PRD_synth_od_kWFs = dict()
for i in range(n_set):
    PRD_synth_od_kWFs[i] = get_PRD(raw_synth_od_cleans[i], raw_synth_od_filtered_kWFs[i], raw_synth_od_unfiltereds[i])

# CC (OD data)
DeltaCC_synth_od_kWFs = dict()
for i in range(n_set):
    DeltaCC_synth_od_kWFs[i] = get_DeltaCC(raw_synth_od_cleans[i], raw_synth_od_filtered_kWFs[i], raw_synth_od_unfiltereds[i])

# CNR Cooper (haemo data)
raw_synth_haemo_filtered_kWFs = dict()
CNR_cooper_synth_haemo_kWFs = dict()
for i in range(n_set):
    raw_synth_haemo_filtered_kWFs[i] = beer_lambert_law(raw_synth_od_filtered_kWFs[i], ppf=ppf)
    CNR_cooper_synth_haemo_kWFs[i] = get_CNR_Cooper(raw_synth_haemo_filtered_kWFs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped

# CNR Zhang (haemo data)
CNR_zhang_synth_haemo_kWFs = dict()
for i in range(n_set):
    CNR_zhang_synth_haemo_kWFs[i] = get_CNR_Zhang(raw_synth_haemo_filtered_kWFs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped

#APA (OD data)
APA_synth_od_kWFs = dict()
for i in range(n_set):
    ids = [event_dict_synths[i]['Spike'], event_dict_synths[i]['Shift']]
    
    start_ind = events_synths[i][np.isin(events_synths[i][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    Am_synth = zip(start_ind, stop_ind)
    
    APA_synth_od_kWFs[i] = get_APA(raw_synth_od_filtered_kWFs[i], raw_synth_od_unfiltereds[i], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

raw_od_filtered_kWF = wavelet_filter_k(raw_od, plot = plot, export = export, filename = 'MA')

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_kWF.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('kbWF/kbWF_MA_detail.png')

S2_D2 760

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_kWF.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('kbWF/kbWF_MA_detail2.png')

S2_D2 760

# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_kWF = get_APA(raw_od_filtered_kWF, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

# Define the MSD function
def MSD(x, k):
    w = 2*k+1
    N = len(x)
    msd = np.zeros_like(x)
    msd[:k] = np.NAN
    msd[-k:] = np.NAN
    for t in range(k,N-k):
        x_window = x[t-k:t+k+1]
        msd[t] = np.sqrt(np.sum(x_window**2)-np.sum(x_window)**2/w)/w
    return msd

# Define a fucntion that detects the start and end points of the MAs
# First segment is lower than threshold!
def detection_start_end(x, k, T = None, n = 3):
    s = MSD(x, k)

    if T == None:
        T = np.nanmean(s) + n * np.nanstd(s)
        
    non_zeros_id = np.where(s > T)[0]
    split = [non_zeros_id[0]]
    for i, id in enumerate(non_zeros_id[1:]):
        if id != (non_zeros_id[i] + 1):
            split.append(int(non_zeros_id[i] + 1))
            split.append(id)
    split.append(non_zeros_id[-1]+1)
    split = list(np.array(split)+k)
    
    x_segmented = np.split(x, split)
    if x_segmented[0].size > 0:
        print('First segment lower than threshold')
        first_element = 0
    else:
        print('First segment higher than threshold')
        s_segmented = s_segmented[1:]
        first_element = 1
    if x_segmented[-1].size > 0:
        print('last segment lower than threshold')
    else:
        print('last segment higher than threshold')
        x_segmented = x_segmented[:-1]
    return x_segmented, first_element

# Define spline interpolation function
# --> each segment containing an MA is spline interpolated

def spline_int(raw, t_start, t_stop, channel, k, T, n = None, p = 0.01, plot = False):
    x = raw.copy().crop(t_start, t_stop).get_data(picks = channel)[0]
    x_org = x.copy()
    x_segmented, first_element = detection_start_end(x, k, T, n)
    
    for i, segment in enumerate(x_segmented):
        if (i%2 != first_element) and (len(segment) > 1):
            x = np.arange(len(segment))
            xs = np.arange(0, len(segment), 0.1)
            yi = csaps(x, segment, xs, smooth = p) # See scholkmann 2010 p = 0.01
            ps = csaps(x, segment, smooth = p)

            corrected_segment = segment - ps(x)
            x_segmented[i] = corrected_segment
        
            if plot:
                plt.plot(x/raw.info['sfreq'], segment, 'o-', label='data')
                plt.plot(xs/raw.info['sfreq'], yi, label="S")
                plt.plot(x/raw.info['sfreq'], corrected_segment, 'x-', label = 'corrected data')
                plt.legend(loc='lower left', ncol=2)
                plt.title('Interpolation of segment ' + str(i))
                plt.show()
    x_corrected = np.concatenate(x_segmented)
    
    if plot:
        t = np.arange(len(x_org))/raw.info['sfreq']
        plt.plot(t, x_org, 'o-', label='data')
        plt.plot(t, x_corrected, 'x-', label = 'corrected data')
        plt.legend(loc='lower left', ncol=2)
        plt.title('Comparison original and MA corrected signal')
        plt.show()
    return x_segmented, first_element

# # W should be similar to the period of spikes that users aimed to identify. see Zhao2021
# t_spike = 1
# samples_spike = t_spike*sfreq
# k = int(samples_spike//2)
# T = None

# x_segmented, first_element = spline_int(raw_synth_od_unfiltered, 0, 499, 1, k, T, n = 2, p = 0.01, plot = True)

# # A lot of coding!! see scholkman 2010
# # Step 4: Reconstruction of the whole time series
# if not first_element:
#     x_segmented_new = []
# else:
#     x_segmented_new = [x_segmented[0]]
# alpha = round(raw_MA_od.info['sfreq']/3)
# beta = round(raw_MA_od.info['sfreq']/2)
# for i in range(0,int(len(x_segmented)-1)):
#     if i%2 == first_element:
#         x_segmented_new.append(x_segmented[i])
#         lambda_1 = len(x_segmented[i])
#         lambda_2 = len(x_segmented[i+1])
#         theta_1 = np.ceil(lambda_1/10)
#         theta_2 = np.ceil(lambda_2/10)
#         if lambda_1 <= alpha:
#             if lambda_2 <= alpha:
#                 a = np.sum(x_segmented[i])
#                 b = np.sum(x_segmented[i+1])
#             elif lambda_2 >= beta:
#                 a = np.sum(x_segmented[i])
#                 b = np.sum(x_segmented[i+1][:int(theta_2+1)])
#             else:
#                 a = np.sum(x_segmented[i][:int(theta_1+1)])
#                 b = np.sum(x_segmented[i+1][:int(alpha+1)])
#         elif lambda_1 >= beta:
#             if lambda_2 <= alpha:
#                 a = np.sum(x_segmented[i][int(lambda_1-theta_1):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:lambda_2])
#             elif lambda_2 >= beta:
#                 a = np.sum(x_segmented[i][int(lambda_1-theta_1):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:int(theta_2+1)])
#             else:
#                 a = np.sum(x_segmented[i][int(lambda_1-theta_1):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:int(alpha+1)])
#         else:
#             if lambda_2 <= alpha:
#                 a = np.sum(x_segmented[i][int(lambda_1-alpha):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:lambda_2])
#             elif lambda_2 >= beta:
#                 a = np.sum(x_segmented[i][int(lambda_1-alpha):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:int(theta_2+1)])
#             else:
#                 a = np.sum(x_segmented[i][int(lambda_1-alpha):int(lambda_1+1)])
#                 b = np.sum(x_segmented[i+1][:int(alpha+1)])
#         theta_m = a - b
#         segment_shifted = x_segmented[i+1] + theta_m
#         x_segmented_new.append(segment_shifted)
# x_new = np.concatenate(x_segmented_new)
# plt.plot(x_new, '--', label = 'filtered')
# plt.plot(x, label = 'orginal')
# plt.legend()
# plt.show()

# Define PCA function
def PCA_Zhang(raw_input, stim_names, DurStim, r = 3, tmin = -5, tmax = 15, plot = False, export = False, filename = None, xlim = None):
    """
    Input:
    - raw_input: raw haemo unfiltered data
    - stim_names: list containing strings which represent names of task stimuli eg: ['tapping/left', 'tapping/right']
    - DurStim: duration of stimuli in seconds
    - r: If 0 < r < 1, select the number of PCA components such that the amount of variance that needs to be explained is greater than the percentage specified by r.
    elif r > 1, number of PCA components to keep
    - tmin, tmax: define duration (in seconds) added around baseline and stimulus segments to perform PCA to eg: 5 seconds before first stimuli and 15 seconds after last stimuli is completed
    """
    raw =raw_input.copy()
    events_, event_dict_ = mne.events_from_annotations(raw)
    fs = raw.info['sfreq']
    
    control_id = event_dict_['Baseline']
    trials_id = []
    for stim in stim_names:
        trials_id.append(event_dict_[stim])

    start_trials = []
    for id in trials_id:
        start_trials += list(events_[events_[:,2] == id][:,0]/raw.info['sfreq'])
    start_trials = np.sort(np.array(start_trials))

    start_controls = []
    start_controls = list(events_[events_[:,2] == control_id][:,0]/raw.info['sfreq'])
    start_controls = np.sort(np.array(start_controls))

    t_start_trials = start_trials[0] + tmin
    t_stop_trials = start_trials[-1] + DurStim + tmax

    t_start_controls = start_controls[0] + tmin
    t_stop_controls = start_controls[-1] + tmax
    
    Z_stim = raw.copy().crop(tmin=t_start_trials, tmax = t_stop_trials).get_data()
    Z_base = raw.copy().crop(tmin=t_start_controls, tmax= t_stop_controls).get_data()

    # removing the mean and scaling to unit variance
    scaler_stim = StandardScaler().fit(Z_stim)
    scaler_base = StandardScaler().fit(Z_base)
    Z_stim = scaler_stim.transform(Z_stim)
    Z_base = scaler_base.transform(Z_base)
    
    if raw.ch_names[0][-3:] == 'hbo':
        Z_base_hbo = Z_base[::2,:]
        Z_base_hbr = Z_base[1::2,:]

        Z_stim_hbo = Z_stim[::2,:]
        Z_stim_hbr = Z_stim[1::2,:]
    else:
        Z_base_hbo = Z_base[1::2,:]
        Z_base_hbr = Z_base[::2,:]

        Z_stim_hbo = Z_stim[1::2,:]
        Z_stim_hbr = Z_stim[::2,:]

    N_base = Z_base.shape[1]
    
    C_base_hbo = np.matmul(Z_base_hbo, Z_base_hbo.T)/N_base
    C_base_hbr = np.matmul(Z_base_hbr, Z_base_hbr.T)/N_base

    eigenvalues_hbo, eigenvectors_hbo = np.linalg.eig(C_base_hbo)
    eigenvalues_hbr, eigenvectors_hbr = np.linalg.eig(C_base_hbr)

    # Sort eigenvalues
    id_hbo = eigenvalues_hbo.argsort()[::-1] 
    eigenvalues_hbo = eigenvalues_hbo[id_hbo]
    eigenvectors_hbo = eigenvectors_hbo[:,id_hbo]
    id_hbr = eigenvalues_hbr.argsort()[::-1]
    eigenvalues_hbr = eigenvalues_hbr[id_hbr]
    eigenvectors_hbr = eigenvectors_hbr[:,id_hbr]
    
    variance_explained_hbo = []
    for eigenvalue in eigenvalues_hbo:
        variance_explained_hbo.append((eigenvalue/sum(eigenvalues_hbo))*100)
    # variance_explained_hbr = []
    # for eigenvalue in eigenvalues_hbr:
    #     variance_explained_hbr.append((eigenvalue/sum(eigenvalues_hbr))*100)

    if r < 1:
        r = 100*r # Convert to percentage
        sum_variance = 0
        for i in range(len(variance_explained_hbo)):
            sum_variance += variance_explained_hbo[i]
            if sum_variance >= r:
                r = i + 1
                print('The PCA is performed with ' + str(r) + ' component(s), representing ' + str(sum_variance) + '% of the variance.')
                break
                
    U_hbo_r = eigenvectors_hbo[:,:r] # Select r first columns corresponding to largest eigenvalues
    U_hbr_r = eigenvectors_hbr[:,:r]

    Z_stim_hbo_new = np.matmul((np.eye(int(len(raw.ch_names)/2)) - np.matmul(U_hbo_r, U_hbo_r.T)), Z_stim_hbo)
    Z_stim_hbr_new = np.matmul((np.eye(int(len(raw.ch_names)/2)) - np.matmul(U_hbr_r, U_hbr_r.T)), Z_stim_hbr)
    
    Z_stim_new = np.zeros_like(Z_stim)
    if raw.ch_names[0][-3:] == 'hbo':
        Z_stim_new[::2,:] = Z_stim_hbo_new
        Z_stim_new[1::2,:] = Z_stim_hbr_new
    else:
        Z_stim_new[1::2,:] = Z_stim_hbo_new
        Z_stim_new[::2,:] = Z_stim_hbr_new
    
    Z_stim_rescaled = scaler_stim.inverse_transform(Z_stim)
    Z_stim_new_rescaled = scaler_stim.inverse_transform(Z_stim_new)
    

    raw_out = raw.copy().crop(tmin=t_start_trials, tmax=t_stop_trials)
    picks = np.sort(_validate_nirs_info(raw_out.info))

    for pick in picks:
        if plot:
            # Plot original and new values
            type = ' hbo' if pick%2 == 0 else ' hbr'
            %matplotlib inline
            time = np.arange(Z_stim_new_rescaled.shape[1])/fs + t_start_trials
            plt.plot(time, Z_stim_rescaled[pick], label = 'original')
            plt.plot(time, Z_stim_new_rescaled[pick], label = 'filtered')
            plt.xlabel('time (s)')
            plt.ylabel('Concentration (M)')
            plt.title('Channel: ' + str(pick//2+1) + type)
            plt.legend()
            if xlim != None:
                plt.xlim(xlim)
            if export:
                if filename == None:
                    raise ValueError('Filename must be given to export figures')
                plt.savefig('PCA/PCA_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
            plt.show()
        
        raw_out._data[pick] = Z_stim_new_rescaled[pick]
    
    return raw_out, t_start_trials, t_stop_trials

t_start_trials, t_stop_trials = dict(), dict()
raw_synth_haemo_filtered_PCAs_2, raw_synth_haemo_filtered_PCAs_3 = dict(), dict()
print('2 components:')
raw_synth_haemo_filtered_PCAs_2[0], t_start_trials[0], t_stop_trials[0] = PCA_Zhang(raw_synth_haemo_unfiltereds[0], ['Task'], stimDur, r = 2, tmin = -5, tmax = 15, plot = plot, export = export, filename = '2_synth', xlim = None)
print('3 components:')
raw_synth_haemo_filtered_PCAs_3[0], _, __ = PCA_Zhang(raw_synth_haemo_unfiltereds[0], ['Task'], stimDur, r = 3, tmin = -5, tmax = 15, plot = plot, export = export, filename = '3_synth', xlim = None)
for i in range(1, n_set):
    raw_synth_haemo_filtered_PCAs_2[i], t_start_trials[i], t_stop_trials[i] = PCA_Zhang(raw_synth_haemo_unfiltereds[i], ['Task'], stimDur, r = 2, tmin = -5, tmax = 15)
    raw_synth_haemo_filtered_PCAs_3[i], t_start_trials[i], t_stop_trials[i] = PCA_Zhang(raw_synth_haemo_unfiltereds[i], ['Task'], stimDur, r = 3, tmin = -5, tmax = 15)

2 components:
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
3 components:
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']

# Convert to OD data to be able to perform all evaluation metrics
raw_synth_od_filtered_PCAs_2 = dict()
raw_synth_od_filtered_PCAs_3 = dict()
for i in range(n_set):
    raw_synth_od_filtered_PCAs_2[i] = inverse_beer_lambert(raw_synth_haemo_filtered_PCAs_2[i], ppf = ppf)
    raw_synth_od_filtered_PCAs_3[i] = inverse_beer_lambert(raw_synth_haemo_filtered_PCAs_3[i], ppf = ppf)

# PCA (r = 2)
# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 185, 215
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_PCAs_2[0].copy().pick(0).crop(tmin - t_start_trials[0],tmax - t_start_trials[0]).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_2_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 315, 345
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_PCAs_2[0].copy().pick(0).crop(tmin - t_start_trials[0],tmax - t_start_trials[0]).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin 
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_2_synth_detail2.png')

# PCA (r = 3)
# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 185, 215
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_PCAs_3[0].copy().pick(0).crop(tmin - t_start_trials[0],tmax - t_start_trials[0]).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_3_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 315, 345
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_PCAs_3[0].copy().pick(0).crop(tmin - t_start_trials[0],tmax - t_start_trials[0]).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin 
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_3_synth_detail2.png')

# # filtered
# X_filtered = raw_synth_haemo_filtered_PCA.get_data().T
# # removing the mean and scaling to unit variance
# x_scaled_filtered = StandardScaler().fit_transform(X_filtered)

# pca_filtered = PCA() 
 
# # Fit and transform data
# pca_features_filtered = pca_filtered.fit_transform(x_scaled_filtered)

# # Unfiltered
# X_unfiltered = raw_synth_haemo_unfiltered.get_data().T
# # removing the mean and scaling to unit variance
# x_scaled_unfiltered = StandardScaler().fit_transform(X_unfiltered)

# pca_unfiltered = PCA()

# # Fit and transform data
# pca_features_unfiltered = pca_unfiltered.fit_transform(x_scaled_unfiltered)

# # Scree Plots
# fig, axes = plt.subplots(1, 2, figsize=(11, 5))

# # Plot for filtered data
# axes[0].bar(range(1, len(pca_filtered.explained_variance_) + 1), pca_filtered.explained_variance_)
# axes[0].plot(range(1, len(pca_filtered.explained_variance_) + 1), np.cumsum(pca_filtered.explained_variance_), c='red', label='Cumulative Explained Variance')
# axes[0].legend(loc='upper left')
# axes[0].set_xlabel('Number of components')
# axes[0].set_ylabel('Explained variance (eigenvalues)')
# axes[0].set_title('Scree Plot - Filtered Data')

# # Plot for unfiltered data
# axes[1].bar(range(1, len(pca_unfiltered.explained_variance_) + 1), pca_unfiltered.explained_variance_)
# axes[1].plot(range(1, len(pca_unfiltered.explained_variance_) + 1), np.cumsum(pca_unfiltered.explained_variance_), c='red', label='Cumulative Explained Variance')
# axes[1].legend(loc='upper left')
# axes[1].set_xlabel('Number of components')
# axes[1].set_ylabel('Explained variance (eigenvalues)')
# axes[1].set_title('Scree Plot - Unfiltered Data')

# # Adjust layout
# plt.tight_layout()
# plt.show()

# SNR (OD data)
DeltaSNR_synth_od_PCAs_2 = dict()
DeltaSNR_synth_od_PCAs_3 = dict()
for i in range(n_set):
    DeltaSNR_synth_od_PCAs_2[i] = get_DeltaSNR(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_2[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))
    DeltaSNR_synth_od_PCAs_3[i] = get_DeltaSNR(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_3[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))

# PRD (OD data)
PRD_synth_od_PCAs_2 = dict()
PRD_synth_od_PCAs_3 = dict()
for i in range(n_set):
    PRD_synth_od_PCAs_2[i] = get_PRD(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_2[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))
    PRD_synth_od_PCAs_3[i] = get_PRD(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_3[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))

# CC (OD data)
DeltaCC_synth_od_PCAs_2 = dict()
DeltaCC_synth_od_PCAs_3 = dict()
for i in range(n_set):
    DeltaCC_synth_od_PCAs_2[i] = get_DeltaCC(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_2[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))
    DeltaCC_synth_od_PCAs_3[i] = get_DeltaCC(raw_synth_od_cleans[i].copy().crop(t_start_trials[i], t_stop_trials[i]), raw_synth_od_filtered_PCAs_3[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]))

# CNR Cooper (haemo data)
CNR_cooper_synth_haemo_PCAs_2 = dict()
CNR_cooper_synth_haemo_PCAs_3 = dict()
for i in range(n_set):
    print(i)
    CNR_cooper_synth_haemo_PCAs_2[i] = get_CNR_Cooper(raw_synth_haemo_filtered_PCAs_2[i], raw_synth_haemo_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)
    CNR_cooper_synth_haemo_PCAs_3[i] = get_CNR_Cooper(raw_synth_haemo_filtered_PCAs_3[i], raw_synth_haemo_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

0
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
4 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
4 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
4 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
4 bad epochs dropped
1
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
7 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
7 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
7 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
7 bad epochs dropped
2
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 212 original time points ...
5 bad epochs dropped
3
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
11 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
11 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
11 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
11 bad epochs dropped
4
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
36 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
36 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
36 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
36 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 212 original time points ...
3 bad epochs dropped
5
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 212 original time points ...
5 bad epochs dropped
6
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 212 original time points ...
5 bad epochs dropped
7
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
2 bad epochs dropped
8
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
39 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
39 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
39 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
39 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 212 original time points ...
3 bad epochs dropped
9
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
7 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
7 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
7 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 212 original time points ...
7 bad epochs dropped

# CNR Zhang (haemo data)
CNR_zhang_synth_haemo_PCAs_2 = dict()
CNR_zhang_synth_haemo_PCAs_3 = dict()
for i in range(n_set):
    CNR_zhang_synth_haemo_PCAs_2[i] = get_CNR_Zhang(raw_synth_haemo_filtered_PCAs_2[i], raw_synth_haemo_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), reject_criteria = reject_criteria)
    CNR_zhang_synth_haemo_PCAs_3[i] = get_CNR_Zhang(raw_synth_haemo_filtered_PCAs_3[i], raw_synth_haemo_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), reject_criteria = reject_criteria)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
7 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
7 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
7 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
7 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
35 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 35 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
9 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
9 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
9 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
9 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
36 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
36 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
36 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
36 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 36 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
38 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 38 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
4 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
4 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
4 bad epochs dropped
Not setting metadata
37 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 37 events and 207 original time points ...
4 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
39 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
39 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
39 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
39 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 39 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
34 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 34 events and 207 original time points ...
5 bad epochs dropped

#APA (OD data)
events_synth_PCAs, event_dict_synth_PCAs = dict(), dict()
APA_synth_od_PCAs_2 = dict()
APA_synth_od_PCAs_3 = dict()

for i in range(n_set):
    # Because length changed, event ids are not correct anymore
    events_synth_PCAs[i], event_dict_synth_PCAs[i] = mne.events_from_annotations(raw_synth_od_filtered_PCAs_2[i])
    
    ids = [event_dict_synth_PCAs[i]['Spike'], event_dict_synth_PCAs[i]['Shift']]
    start_ind = events_synth_PCAs[i][np.isin(events_synth_PCAs[i][:,2], ids)][:,0] - events_synth_PCAs[i][0,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    
    Am_synth = zip(start_ind, stop_ind)
    
    APA_synth_od_PCAs_2[i] = get_APA(raw_synth_od_filtered_PCAs_2[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)
    APA_synth_od_PCAs_3[i] = get_APA(raw_synth_od_filtered_PCAs_3[i], raw_synth_od_unfiltereds[i].copy().crop(t_start_trials[i], t_stop_trials[i]), Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Used Annotations descriptions: ['Shift', 'Spike', 'Task']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

raw_haemo_filtered_PCA_2, t_start_trials, t_stop_trials = PCA_Zhang(raw_haemo, ['test/Horizontal', 'test/Random', 'test/Vertical'], 3, r = 2, tmin = -5, tmax = 15, plot = plot, export = export, filename = '2_MA', xlim = None)
raw_haemo_filtered_PCA_3, t_start_trials, t_stop_trials = PCA_Zhang(raw_haemo, ['test/Horizontal', 'test/Random', 'test/Vertical'], 3, r = 3, tmin = -5, tmax = 15, plot = plot, export = export, filename = '3_MA', xlim = None)

Used Annotations descriptions: ['Baseline', 'practice/Horizontal', 'practice/Random', 'practice/Vertical', 'test/Horizontal', 'test/Random', 'test/Vertical']
Used Annotations descriptions: ['Baseline', 'practice/Horizontal', 'practice/Random', 'practice/Vertical', 'test/Horizontal', 'test/Random', 'test/Vertical']

# Convert to OD data
raw_od_filtered_PCA_2 = inverse_beer_lambert(raw_haemo_filtered_PCA_2, ppf = ppf)
raw_od_filtered_PCA_3 = inverse_beer_lambert(raw_haemo_filtered_PCA_3, ppf = ppf)

# PCA (r=2)
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_PCA_2.copy().pick(chan).crop(tmin-t_start_trials,tmax-t_start_trials).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_2_MA_detail.png')

S2_D2 760

# PCA (r=2)
%matplotlib inline
chan = 22
print(raw_od.ch_names[chan])
tmin, tmax = 355, 385
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_PCA_2.copy().pick(chan).crop(tmin-t_start_trials,tmax-t_start_trials).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S8-D4 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_2_MA_detail2.png')

S8_D4 760

# PCA (r=3)
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_PCA_3.copy().pick(chan).crop(tmin-t_start_trials,tmax-t_start_trials).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_3_MA_detail.png')

S2_D2 760

# PCA (r=3)
%matplotlib inline
chan = 22
print(raw_od.ch_names[chan])
tmin, tmax = 355, 385
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_PCA_3.copy().pick(chan).crop(tmin-t_start_trials,tmax-t_start_trials).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
t2 = np.arange(len(data_filtered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t2, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S8-D4 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('PCA/PCA_3_MA_detail2.png')

S8_D4 760

#APA (OD data)

# Because length changed, event ids are not correct anymore
events_PCA, event_dict_PCA = mne.events_from_annotations(raw_haemo_filtered_PCA_2)

ids = [event_dict_PCA['test/Horizontal'], event_dict_PCA['test/Random'], event_dict_PCA['test/Vertical']]
start_ind = events_PCA[np.isin(events_PCA[:,2], ids)][:,0] - events_PCA[0,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact

Am = zip(start_ind, stop_ind)

APA_od_PCA_2 = get_APA(raw_od_filtered_PCA_2, raw_od.copy().crop(t_start_trials, t_stop_trials), Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)
APA_od_PCA_3 = get_APA(raw_od_filtered_PCA_3, raw_od.copy().crop(t_start_trials, t_stop_trials), Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Used Annotations descriptions: ['test/Horizontal', 'test/Random', 'test/Vertical']
Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

# Using ICA interface in MNE-Python (similar to the interface in scikit-learn)
def ICA_nirs(raw_input, n_components = 0.99, method = 'fastica', score_method = 'MSD', IMU_df = None, plot = False, xlim = None, export = False, filename = None):
    
    raw = raw_input.copy()
    fs = raw.info['sfreq']
    data = raw.get_data()
    
    # 'Before we run the ICA, an important step is filtering the data to remove low-frequency drifts'
    # --> Not feasable as important information is under 1 Hz

    # Scaling to unit variance and whitened using PCA 
    ica = ICA(n_components = n_components, method = method) 
    ica.fit(raw)
    print(ica)

    # # Plot ICA sources
    # if plot:
    #     %matplotlib qt
    #     ica.plot_sources(raw, show_scrollbars = True);

    # Get sources data for score calculation
    ica_sources = ica.get_sources(raw).get_data()
    
    if score_method == 'MSD':
        # Perform MSD
        # W = 2*k + 1 should be similar to the period of spikes that users aimed to identify. See Zhao et al. (2021)
        t_spike = 1
        samples_spike = t_spike*raw.info['sfreq']
        k = int(samples_spike//2)
        n = 3 # Threshold 3 std from the mean
        
        scores = []
        for i in range(ica_sources.shape[0]):
            s = MSD(ica_sources[i], k)
            T = np.nanmean(s) + n*np.nanstd(s)
            score = np.sum(np.where(s > T, 1, 0))
            scores.append(score)
        scores = np.array(scores)

    elif score_method == 'CEI':
        # Perform correlation with external input
        # Need IMU data  
        IMU = IMU_df.to_numpy().T
        scores = []
        for i in range(ica_sources.shape[0]):
            corr_xyz = []
            for j in range(6): # x, y and z acc + 3 gyro recorded
                corr = pearsonr(ica_sources[i], IMU[j])
                corr_xyz.append(corr)
            # Take mean?
            score = np.mean(np.array(corr_xyz))
            scores.append(score)

    else:
        raise ValueError('Please add scoring method')
        
    if plot: 
        # Plot scores
        %matplotlib inline
        fig = ica.plot_scores(scores)
        if export:
            if filename == None:
                raise ValueError('Filename must be given to export figures')
            fig.savefig('ICA/ICA_scores' + filename + '.png')
    
    # Remove source with highest score
    ica.exclude = [np.argmax(np.array(scores))]

    # Reconstruct to get clean fNIRS
    # ICA apply() does:
        # 1) Unmixes the data with the unmixing_matrix
        # 2) Includes ICA components based on ica.include and ica.exclude
        # 3) Re-mixes the data with mixing_matrix_
        # 4) Restores any data not passed to the ICA algorithm, i.e., the PCA components between n_components and n_pca_components
    reconst_raw = raw.copy()
    ica.apply(reconst_raw)
    data_new = reconst_raw.get_data()
    
    picks = np.sort(_validate_nirs_info(raw.info))
    datatype = raw.ch_names[0][-3:]
    
    for pick in picks:
        if plot:
            # Plot original and new values
            if datatype == 'hbo':
                type = ' hbo' if pick%2 == 0 else ' hbr'
                ylabel = 'HbO/HbR (M)'
            else:
                type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
                ylabel = 'OD (V)'
            %matplotlib inline
            time = np.arange(data.shape[1])/fs
            plt.plot(time, data[pick], label = 'original')
            plt.plot(time, data_new[pick], label = 'filtered')
            plt.xlabel('time (s)')
            plt.ylabel(ylabel)
            plt.title('Channel: ' + str(pick//2+1) + type)
            plt.legend()
            if xlim != None:
                plt.xlim(xlim)
            if export:
                if filename == None:
                    raise ValueError('Filename must be given to export figures')
                plt.savefig('ICA/ICA_' + score_method + '_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
            plt.show()

    return reconst_raw

raw_synth_od_filtered_ICAs = dict()
raw_synth_od_filtered_ICAs[0] = ICA_nirs(raw_synth_od_unfiltereds[0], n_components = 0.99, method = 'fastica', score_method = 'MSD', IMU_df = None, plot = plot, xlim = None, export = export, filename = 'synth')
for i in range(1, n_set):
    raw_synth_od_filtered_ICAs[i] = ICA_nirs(raw_synth_od_unfiltereds[i], n_components = 0.99, method = 'fastica', score_method = 'MSD')

Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.0s.
<ICA | raw data decomposition, method: fastica (fit in 17 iterations on 2940 samples), 5 ICA components (8 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component
    Projecting back using 8 PCA components
Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.0s.
<ICA | raw data decomposition, method: fastica (fit in 8 iterations on 2940 samples), 5 ICA components (8 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance
    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component
    Projecting back using 8 PCA components
Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.0s.

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)
C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

<ICA | raw data decomposition, method: fastica (fit in 22 iterations on 2940 samples), 5 ICA components (8 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance
    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component
    Projecting back using 8 PCA components
Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.0s.
<ICA | raw data decomposition, method: fastica (fit in 10 iterations on 2940 samples), 5 ICA components (8 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance
    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

    Projecting back using 8 PCA components
Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 4 components

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

Fitting ICA took 0.3s.
<ICA | raw data decomposition, method: fastica (fit in 1000 iterations on 2940 samples), 4 ICA components (8 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance
    Transforming to ICA space (4 components)
    Zeroing out 1 ICA component
    Projecting back using 8 PCA components
Fitting ICA to data using 8 channels (please be patient, this may take a while)
Selecting by explained variance: 3 components
Fitting ICA took 0.0s.

C:\Users\fabia\AppData\Roaming\Python\Python311\site-packages\sklearn\decomposition\_fastica.py:128: ConvergenceWarning: FastICA did not converge. Consider increasing tolerance or the maximum number of iterations.
  warnings.warn(
C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 10, 40
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_ICAs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 310, 340
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_ICAs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq +tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_synth_detail2.png')

# SNR (OD data)
DeltaSNR_synth_od_ICAs = dict()
for i in range(n_set):
    DeltaSNR_synth_od_ICAs[i] = get_DeltaSNR(raw_synth_od_cleans[i], raw_synth_od_filtered_ICAs[i], raw_synth_od_unfiltereds[i])

# PRD (OD data)
PRD_synth_od_ICAs = dict()
for i in range(n_set):
    PRD_synth_od_ICAs[i] = get_PRD(raw_synth_od_cleans[i], raw_synth_od_filtered_ICAs[i], raw_synth_od_unfiltereds[i])

# CC (OD data)
DeltaCC_synth_od_ICAs = dict()
for i in range(n_set):
    DeltaCC_synth_od_ICAs[i] = get_DeltaCC(raw_synth_od_cleans[i], raw_synth_od_filtered_ICAs[i], raw_synth_od_unfiltereds[i])

# CNR Cooper (haemo data)
raw_synth_haemo_filtered_ICAs = dict()
CNR_cooper_synth_haemo_ICAs = dict()
for i in range(n_set):
    raw_synth_haemo_filtered_ICAs[i] = beer_lambert_law(raw_synth_od_filtered_ICAs[i], ppf = ppf)
    CNR_cooper_synth_haemo_ICAs[i] = get_CNR_Cooper(raw_synth_haemo_filtered_ICAs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped

# CNR Zhang (haemo data)
CNR_zhang_synth_haemo_ICAs = dict()
for i in range(n_set):
    CNR_zhang_synth_haemo_ICAs[i] = get_CNR_Zhang(raw_synth_haemo_filtered_ICAs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped

#APA (OD data)
APA_synth_od_ICAs = dict()
for i in range(n_set):
    ids = [event_dict_synths[i]['Spike'], event_dict_synths[i]['Shift']]
    
    start_ind = events_synths[i][np.isin(events_synths[i][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    Am_synth = zip(start_ind, stop_ind)
    
    APA_synth_od_ICAs[i] = get_APA(raw_synth_od_filtered_ICAs[i], raw_synth_od_unfiltereds[i], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

# MSD
raw_od_filtered_ICA_MSD = ICA_nirs(raw_od, n_components = 0.99, method = 'fastica', score_method = 'MSD', IMU_df = None, plot = plot, xlim = None, export = export, filename = 'MA')

Fitting ICA to data using 52 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.1s.
<ICA | raw data decomposition, method: fastica (fit in 13 iterations on 5163 samples), 5 ICA components (52 PCA components available), channel types: fnirs_od, no sources marked for exclusion>

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

Applying ICA to Raw instance
    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component
    Projecting back using 52 PCA components

# CEI
raw_od_filtered_ICA_CEI = ICA_nirs(raw_od, n_components = 0.99, method = 'fastica', score_method = 'CEI', IMU_df = IMU_data, plot = plot, xlim = None, export = export, filename = 'MA')

Fitting ICA to data using 52 channels (please be patient, this may take a while)
Selecting by explained variance: 5 components
Fitting ICA took 0.0s.
<ICA | raw data decomposition, method: fastica (fit in 12 iterations on 5163 samples), 5 ICA components (52 PCA components available), channel types: fnirs_od, no sources marked for exclusion>
Applying ICA to Raw instance
    Transforming to ICA space (5 components)
    Zeroing out 1 ICA component
    Projecting back using 52 PCA components

C:\Users\fabia\AppData\Local\Temp\ipykernel_27480\1915246416.py:13: RuntimeWarning: The data has not been high-pass filtered. For good ICA performance, it should be high-pass filtered (e.g., with a 1.0 Hz lower bound) before fitting ICA.
  ica.fit(raw)

# MSD
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_ICA_MSD.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_MSD_MA_detail.png')

S2_D2 760

# MSD
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_ICA_MSD.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_MSD_MA_detail2.png')

S2_D2 760

# CEI
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_ICA_CEI.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_CEI_MA_detail.png')

S2_D2 760

# CEI
%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_ICA_CEI.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('ICA/ICA_CEI_MA_detail2.png')

S2_D2 760

# MSD
# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_ICA_MSD = get_APA(raw_od_filtered_ICA_MSD, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

# CEI
# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_ICA_CEI = get_APA(raw_od_filtered_ICA_CEI, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

# This approach corrects baseline shift and spike artifacts without the need for any user-supplied parameters (Fishburn Et Al., 2019)
# doi:10.1016/j.neuroimage.2018.09.025
raw_synth_od_filtered_tddrs = dict()
for i in range(n_set):
    raw_synth_od_filtered_tddrs[i] = temporal_derivative_distribution_repair(raw_synth_od_unfiltereds[i])

# Should have made a function for this instead of pasting it in each method
data_new = raw_synth_od_filtered_tddrs[0].get_data()
data = raw_synth_od_unfiltereds[0].get_data()
picks = np.sort(_validate_nirs_info(raw_synth_od_filtered_tddrs[0].info))
filename = 'synth'
for pick in picks:
    if plot:
        type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
        ylabel = 'OD (V)'
        %matplotlib inline
        time = np.arange(data.shape[1])/sfreq
        plt.plot(time, data[pick], label = 'original')
        plt.plot(time, data_new[pick], label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            if filename == None:
                raise ValueError('Filename must be given to export figures')
            plt.savefig('tddr/tddr_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 10, 40
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_tddrs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('tddr/tddr_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 310, 340
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_tddrs[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq +tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('tddr/tddr_synth_detail2.png')

# SNR (OD data)
DeltaSNR_synth_od_tddrs = dict()
for i in range(n_set):
    DeltaSNR_synth_od_tddrs[i] = get_DeltaSNR(raw_synth_od_cleans[i], raw_synth_od_filtered_tddrs[i], raw_synth_od_unfiltereds[i])

# PRD (OD data)
PRD_synth_od_tddrs = dict()
for i in range(n_set):
    PRD_synth_od_tddrs[i] = get_PRD(raw_synth_od_cleans[i], raw_synth_od_filtered_tddrs[i], raw_synth_od_unfiltereds[i])

# CC (OD data)
DeltaCC_synth_od_tddrs = dict()
for i in range(n_set):
    DeltaCC_synth_od_tddrs[i] = get_DeltaCC(raw_synth_od_cleans[i], raw_synth_od_filtered_tddrs[i], raw_synth_od_unfiltereds[i])

# CNR Cooper (haemo data)
raw_synth_haemo_filtered_tddrs = dict()
CNR_cooper_synth_haemo_tddrs = dict()
for i in range(n_set):
    raw_synth_haemo_filtered_tddrs[i] = beer_lambert_law(raw_synth_od_filtered_tddrs[i], ppf=ppf)
    CNR_cooper_synth_haemo_tddrs[i] = get_CNR_Cooper(raw_synth_haemo_filtered_tddrs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped

# CNR Zhang (OD data)
CNR_zhang_synth_haemo_tddrs = dict()
for i in range(n_set):
    CNR_zhang_synth_haemo_tddrs[i] = get_CNR_Zhang(raw_synth_haemo_filtered_tddrs[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped

#APA (OD data)
APA_synth_od_tddrs = dict()
for i in range(n_set):
    ids = [event_dict_synths[i]['Spike'], event_dict_synths[i]['Shift']]
    
    start_ind = events_synths[i][np.isin(events_synths[i][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    Am_synth = zip(start_ind, stop_ind)
    
    APA_synth_od_tddrs[i] = get_APA(raw_synth_od_filtered_tddrs[i], raw_synth_od_unfiltereds[i], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

raw_od_filtered_tddr =  temporal_derivative_distribution_repair(raw_od)

data_new = raw_od_filtered_tddr.get_data()
data = raw_od.get_data()
picks = np.sort(_validate_nirs_info(raw_od.info))
filename = 'MA'
for pick in picks:
    if plot:
        type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
        ylabel = 'OD (V)'
        %matplotlib inline
        time = np.arange(data.shape[1])/sfreq
        plt.plot(time, data[pick], label = 'original')
        plt.plot(time, data_new[pick], label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            if filename == None:
                raise ValueError('Filename must be given to export figures')
            plt.savefig('tddr/tddr_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_tddr.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('tddr/tddr_MA_detail.png')

S2_D2 760

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_tddr.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('tddr/tddr_MA_detail2.png')

S2_D2 760

# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_tddr = get_APA(raw_od_filtered_tddr, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

raw_synth_haemo_filtered_cbsis = dict()
for i in range(n_set):
    raw_synth_haemo_filtered_cbsis[i] = enhance_negative_correlation(raw_synth_haemo_unfiltereds[i])

raw_synth_od_filtered_cbsis = dict()
for i in range(n_set):
    raw_synth_od_filtered_cbsis[i] = inverse_beer_lambert(raw_synth_haemo_filtered_cbsis[i], ppf = ppf)

# Should have made a function for this instead of pasting it in each method
data_new = raw_synth_od_filtered_cbsis[0].get_data()
data = raw_synth_od_unfiltereds[0].get_data()
picks = np.sort(_validate_nirs_info(raw_synth_od_filtered_cbsis[0].info))
filename = 'synth'
for pick in picks:
    if plot:
        type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
        ylabel = 'OD (V)'
        %matplotlib inline
        time = np.arange(data.shape[1])/sfreq
        plt.plot(time, data[pick], label = 'original')
        plt.plot(time, data_new[pick], label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            if filename == None:
                raise ValueError('Filename must be given to export figures')
            plt.savefig('cbsi/cbsi_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

# Zoom in on spike of first dataset (Look for good segment in full recording first)
%matplotlib inline
tmin, tmax = 10, 40
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_cbsis[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq + tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('cbsi/cbsi_synth_detail.png')

# Zoom in on baseline shift of first data
%matplotlib inline
tmin, tmax = 310, 340
data_clean = raw_synth_od_cleans[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_unfiltered = raw_synth_od_unfiltereds[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
data_filtered = raw_synth_od_filtered_cbsis[0].copy().pick(0).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_clean))/sfreq +tmin
plt.plot(t, data_clean, label = 'Motion artefact free', color = 'b')
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S1-D1 760 of synthetic data')
plt.legend()
if export:
    plt.savefig('cbsi/cbsi_synth_detail2.png')

# SNR (OD data)
DeltaSNR_synth_od_cbsis = dict()
for i in range(n_set):
    DeltaSNR_synth_od_cbsis[i] = get_DeltaSNR(raw_synth_od_cleans[i], raw_synth_od_filtered_cbsis[i], raw_synth_od_unfiltereds[i])

# PRD (OD data)
PRD_synth_od_cbsis = dict()
for i in range(n_set):
    PRD_synth_od_cbsis[i] = get_PRD(raw_synth_od_cleans[i], raw_synth_od_filtered_cbsis[i], raw_synth_od_unfiltereds[i])

# CC (OD data)
DeltaCC_synth_od_cbsis = dict()
for i in range(n_set):
    DeltaCC_synth_od_cbsis[i] = get_DeltaCC(raw_synth_od_cleans[i], raw_synth_od_filtered_cbsis[i], raw_synth_od_unfiltereds[i])

# CNR Cooper(haemo data)
CNR_cooper_synth_haemo_cbsis = dict()
for i in range(n_set):
    CNR_cooper_synth_haemo_cbsis[i] = get_CNR_Cooper(raw_synth_haemo_filtered_cbsis[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria, tmin = tmin_CNR, tmax = tmax_CNR, t_peak = t_peak, real_HRF = real_HRF)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-5.9523809523809526, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 212 original time points ...
3 bad epochs dropped

# CNR Zhang (haemo data)
CNR_zhang_synth_haemo_cbsis = dict()
for i in range(n_set):
    CNR_zhang_synth_haemo_cbsis[i] = get_CNR_Zhang(raw_synth_haemo_filtered_cbsis[i], raw_synth_haemo_unfiltereds[i], reject_criteria = reject_criteria)

Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
5 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
2 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
1 bad epochs dropped
Used Annotations descriptions: ['Baseline', 'Shift', 'Spike', 'Task']
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped
Not setting metadata
53 matching events found
Setting baseline interval to [-14.965986394557824, 0.0] s
Applying baseline correction (mode: mean)
0 projection items activated
Using data from preloaded Raw for 53 events and 207 original time points ...
3 bad epochs dropped

# APA (OD data)
APA_synth_od_cbsis = dict()
for i in range(n_set):
    ids = [event_dict_synths[i]['Spike'], event_dict_synths[i]['Shift']]
    start_ind = events_synths[i][np.isin(events_synths[i][:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
    start_ind = np.where(start_ind < 0, 0, start_ind)
    stop_ind = start_ind + int(10*sfreq) # 5 seconds after artefact
    Am_synth = zip(start_ind, stop_ind)
    
    APA_synth_od_cbsis[i] = get_APA(raw_synth_od_filtered_cbsis[i], raw_synth_od_unfiltereds[i], Am_synth, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.054 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s finished

raw_haemo_filtered_cbsi =  enhance_negative_correlation(raw_haemo)

raw_od_filtered_cbsi = inverse_beer_lambert(raw_haemo_filtered_cbsi, ppf = ppf)

# Should have made a function for this instead of pasting it in each method
data_new = raw_od_filtered_cbsi.get_data()
data = raw_od.get_data()
picks = np.sort(_validate_nirs_info(raw_od.info))
filename = 'MA'
for pick in picks:
    if plot:
        type = ' wavelength 760' if pick%2 == 0 else ' wavelength 850'
        ylabel = 'OD (V)'
        %matplotlib inline
        time = np.arange(data.shape[1])/sfreq
        plt.plot(time, data[pick], label = 'original')
        plt.plot(time, data_new[pick], label = 'filtered')
        plt.xlabel('time (s)')
        plt.ylabel(ylabel)
        plt.title('Channel: ' + str(pick//2+1) + type)
        plt.legend()
        if export:
            if filename == None:
                raise ValueError('Filename must be given to export figures')
            plt.savefig('cbsi/cbsi_' + filename + '_channel_' + str(pick//2+1) + type + '.png')
        plt.show()

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 315, 345
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_cbsi.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('cbsi/cbsi_MA_detail.png')

S2_D2 760

%matplotlib inline
chan = 2
print(raw_od.ch_names[chan])
tmin, tmax = 120, 150
data_unfiltered = raw_od.copy().pick(chan).crop(tmin,tmax).get_data()[0]
data_filtered = raw_od_filtered_cbsi.copy().pick(chan).crop(tmin,tmax).get_data()[0]
t = np.arange(len(data_unfiltered))/sfreq + tmin
plt.plot(t, data_unfiltered, label = 'unfiltered', color = 'r')
plt.plot(t, data_filtered, label = 'filtered', color = 'g')
plt.xlabel('time (s)')
plt.ylabel('OD (V)')
plt.title('Detailed view of OD channel S2-D2 760 of MA experiment data')
plt.legend()
if export:
    plt.savefig('cbsi/cbsi_MA_detail2.png')

S2_D2 760

# APA (OD data)
ids = [event_dict['test/Horizontal'], event_dict['test/Random'], event_dict['test/Vertical']]
start_ind = events[np.isin(events[:,2], ids)][:,0] - int(5*sfreq) # 5 seconds before artefact
start_ind = np.where(start_ind < 0, 0, start_ind)
stop_ind = start_ind + int(15*sfreq) # 10 seconds after artefact
Am = zip(start_ind, stop_ind)

APA_od_cbsi = get_APA(raw_od_filtered_cbsi, raw_od, Am, l_freq, l_trans_bandwidth = l_trans_bandwidth)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

Filtering raw data in 1 contiguous segment
Setting up high-pass filter at 1.5 Hz

FIR filter parameters
---------------------
Designing a one-pass, zero-phase, non-causal highpass filter:
- Windowed time-domain design (firwin) method
- Hamming window with 0.0194 passband ripple and 53 dB stopband attenuation
- Lower passband edge: 1.50
- Lower transition bandwidth: 0.30 Hz (-6 dB cutoff frequency: 1.35 Hz)
- Filter length: 65 samples (11.050 s)

[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[Parallel(n_jobs=1)]: Done  52 out of  52 | elapsed:    0.0s finished

metrics = ['DeltaSNR', 'PRD', 'DeltaCC', 'CNR_C', 'CNR_Z', 'APA']
methods = ['WF', 'kbWF', 'PCA_2', 'PCA_3', 'ICA', 'tddr', 'cbsi']
channels = raw_synth_od_cleans[0].ch_names

DeltaSNR_synth_od_WF = pd.DataFrame(DeltaSNR_synth_od_WFs.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_kWF = pd.DataFrame(DeltaSNR_synth_od_kWFs.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_PCA_2 = pd.DataFrame(DeltaSNR_synth_od_PCAs_2.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_PCA_3 = pd.DataFrame(DeltaSNR_synth_od_PCAs_3.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_ICA = pd.DataFrame(DeltaSNR_synth_od_ICAs.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_tddr = pd.DataFrame(DeltaSNR_synth_od_tddrs.values(), columns = channels).mean(axis = 0)
DeltaSNR_synth_od_cbsi = pd.DataFrame(DeltaSNR_synth_od_cbsis.values(), columns = channels).mean(axis = 0)
DeltaSNR = pd.concat([DeltaSNR_synth_od_WF, DeltaSNR_synth_od_kWF, DeltaSNR_synth_od_PCA_2, DeltaSNR_synth_od_PCA_3, DeltaSNR_synth_od_ICA, DeltaSNR_synth_od_tddr, DeltaSNR_synth_od_cbsi], axis = 1)
DeltaSNR.columns = methods
DeltaSNR.plot.bar(title = 'DeltaSNR scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'SNR (dB)', xlabel = 'Channels').get_figure().savefig('DeltaSNR.png', bbox_inches = 'tight')

# Mean score over all channels
scores_SNR = DeltaSNR.mean(axis = 0).sort_values(ascending = False)
scores_SNR

tddr     14.063028
cbsi     11.192195
ICA       1.872284
WF        0.474114
kbWF      0.122899
PCA_3    -0.524182
PCA_2    -0.653897
dtype: float64

print('The best MA correction technique based on SNR value is: ' + DeltaSNR.mean(axis = 0).idxmax() + ' (value = ' + str(DeltaSNR.mean(axis = 0).max()) + ')')

The best MA correction technique based on SNR value is: tddr (value = 14.06302847839465)

PRD_synth_od_WF = pd.DataFrame(PRD_synth_od_WFs.values(), columns = channels).mean(axis = 0)
PRD_synth_od_kWF = pd.DataFrame(PRD_synth_od_kWFs.values(), columns = channels).mean(axis = 0)
PRD_synth_od_PCA_2 = pd.DataFrame(PRD_synth_od_PCAs_2.values(), columns = channels).mean(axis = 0)
PRD_synth_od_PCA_3 = pd.DataFrame(PRD_synth_od_PCAs_3.values(), columns = channels).mean(axis = 0)
PRD_synth_od_ICA = pd.DataFrame(PRD_synth_od_ICAs.values(), columns = channels).mean(axis = 0)
PRD_synth_od_tddr = pd.DataFrame(PRD_synth_od_tddrs.values(), columns = channels).mean(axis = 0)
PRD_synth_od_cbsi = pd.DataFrame(PRD_synth_od_cbsis.values(), columns = channels).mean(axis = 0)
PRD = pd.concat([PRD_synth_od_WF, PRD_synth_od_kWF, PRD_synth_od_PCA_2, PRD_synth_od_PCA_3, PRD_synth_od_ICA, PRD_synth_od_tddr, PRD_synth_od_cbsi], axis = 1)
PRD.columns = methods
PRD.plot.bar(title = 'PRD scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'PRD (%)', xlabel = 'Channels').get_figure().savefig('PRD.png', bbox_inches = 'tight')

# Mean score over all channels
scores_PRD = PRD.mean(axis = 0).sort_values(ascending = True)
scores_PRD

tddr     127.054428
cbsi     212.311875
ICA      548.468075
WF       647.626967
PCA_3    662.438149
kbWF     668.091786
PCA_2    672.737284
dtype: float64

print('The best MA correction technique based on PRD value is: ' + PRD.mean(axis = 0).idxmin() + ' (value = ' + str(PRD.mean(axis = 0).min()) + ')')

The best MA correction technique based on PRD value is: tddr (value = 127.05442774617654)

DeltaCC_synth_od_WF = pd.DataFrame(DeltaCC_synth_od_WFs.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_kWF = pd.DataFrame(DeltaCC_synth_od_kWFs.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_PCA_2 = pd.DataFrame(DeltaCC_synth_od_PCAs_2.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_PCA_3 = pd.DataFrame(DeltaCC_synth_od_PCAs_3.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_ICA = pd.DataFrame(DeltaCC_synth_od_ICAs.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_tddr = pd.DataFrame(DeltaCC_synth_od_tddrs.values(), columns = channels).mean(axis = 0)
DeltaCC_synth_od_cbsi = pd.DataFrame(DeltaCC_synth_od_cbsis.values(), columns = channels).mean(axis = 0)
DeltaCC = pd.concat([DeltaCC_synth_od_WF, DeltaCC_synth_od_kWF, DeltaCC_synth_od_PCA_2, DeltaCC_synth_od_PCA_3, DeltaCC_synth_od_ICA, DeltaCC_synth_od_tddr, DeltaCC_synth_od_cbsi], axis = 1)
DeltaCC.columns = methods

DeltaCC.plot.bar(title = 'DeltaCC scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'DeltaCC', xlabel = 'Channels').get_figure().savefig('DeltaCC.png', bbox_inches = 'tight')

# Mean score over all channels
scores_CC = DeltaCC.mean(axis = 0).sort_values(ascending = False)
scores_CC

tddr     0.271757
ICA      0.025733
kbWF    -0.002573
PCA_2   -0.019410
WF      -0.032031
PCA_3   -0.035556
cbsi    -0.063097
dtype: float64

print('The best MA correction technique based on DeltaCC value is: ' + DeltaCC.mean(axis = 0).idxmax() + ' (value = ' + str(DeltaCC.mean(axis = 0).max()) + ')')

The best MA correction technique based on DeltaCC value is: tddr (value = 0.2717567910358382)

channels = []
for chan in raw_synth_od_cleans[0].ch_names[::2]:
    channels.append(chan[:-3])
channels

['S1_D1 ', 'S3_D1 ', 'S5_D1 ', 'S7_D1 ']

CNR_C_synth_haemo_WF = pd.DataFrame(CNR_cooper_synth_haemo_WFs.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_kWF = pd.DataFrame(CNR_cooper_synth_haemo_kWFs.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_PCA_2 = pd.DataFrame(CNR_cooper_synth_haemo_PCAs_2.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_PCA_3 = pd.DataFrame(CNR_cooper_synth_haemo_PCAs_3.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_ICA = pd.DataFrame(CNR_cooper_synth_haemo_ICAs.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_tddr = pd.DataFrame(CNR_cooper_synth_haemo_tddrs.values(), columns = channels).mean(axis = 0)
CNR_C_synth_haemo_cbsi = pd.DataFrame(CNR_cooper_synth_haemo_cbsis.values(), columns = channels).mean(axis = 0)
CNR_C = pd.concat([CNR_C_synth_haemo_WF, CNR_C_synth_haemo_kWF, CNR_C_synth_haemo_PCA_2, CNR_C_synth_haemo_PCA_3, CNR_C_synth_haemo_ICA, CNR_C_synth_haemo_tddr, CNR_C_synth_haemo_cbsi], axis = 1)
CNR_C.columns = methods

CNR_C.plot.bar(title = 'CNR (Cooper) scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'CNR', xlabel = 'Channels').get_figure().savefig('CNR_C.png', bbox_inches = 'tight')

# Mean score over all channels
scores_CNR_C = CNR_C.mean(axis = 0).sort_values(ascending = False)
scores_CNR_C

cbsi     4.454349
ICA      3.522182
tddr     3.050674
kbWF     2.508243
PCA_2    2.409710
WF       1.624555
PCA_3    1.521516
dtype: float64

print('The best MA correction technique based on CNR (Cooper) value is: ' + CNR_C.mean(axis = 0).idxmax() + ' (value = ' + str(CNR_C.mean(axis = 0).max()) + ')')

The best MA correction technique based on CNR (Cooper) value is: cbsi (value = 4.454349318623494)

CNR_Z_synth_haemo_WF = pd.DataFrame(CNR_zhang_synth_haemo_WFs.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_kWF = pd.DataFrame(CNR_zhang_synth_haemo_kWFs.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_PCA_2 = pd.DataFrame(CNR_zhang_synth_haemo_PCAs_2.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_PCA_3 = pd.DataFrame(CNR_zhang_synth_haemo_PCAs_3.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_ICA = pd.DataFrame(CNR_zhang_synth_haemo_ICAs.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_tddr = pd.DataFrame(CNR_zhang_synth_haemo_tddrs.values(), columns = channels).mean(axis = 0)
CNR_Z_synth_haemo_cbsi = pd.DataFrame(CNR_zhang_synth_haemo_cbsis.values(), columns = channels).mean(axis = 0)
CNR_Z = pd.concat([CNR_Z_synth_haemo_WF, CNR_Z_synth_haemo_kWF, CNR_Z_synth_haemo_PCA_2, CNR_Z_synth_haemo_PCA_3, CNR_Z_synth_haemo_ICA, CNR_Z_synth_haemo_tddr, CNR_Z_synth_haemo_cbsi], axis = 1)
CNR_Z.columns = methods

CNR_Z.plot.bar(title = 'CNR (Zhang) scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'CNR', xlabel = 'Channels').get_figure().savefig('CNR_Z.png', bbox_inches = 'tight')

# Mean score over all channels
scores_CNR_Z = CNR_Z.mean(axis = 0).sort_values(ascending = False)
scores_CNR_Z

PCA_3    0.000283
PCA_2    0.000143
WF      -0.001104
tddr    -0.001558
ICA     -0.001978
cbsi    -0.002464
kbWF    -0.002698
dtype: float64

print('The best MA correction technique based on CNR (Zhang) value is: ' + CNR_Z.mean(axis = 0).idxmax() + ' (value = ' + str(CNR_Z.mean(axis = 0).max()) + ')')

The best MA correction technique based on CNR (Zhang) value is: PCA_3 (value = 0.00028261888314476386)

channels = raw_synth_od_cleans[0].ch_names

APA_synth_od_WF = pd.DataFrame(APA_synth_od_WFs.values(), columns = channels).mean(axis = 0)
APA_synth_od_kWF = pd.DataFrame(APA_synth_od_kWFs.values(), columns = channels).mean(axis = 0)
APA_synth_od_PCA_2 = pd.DataFrame(APA_synth_od_PCAs_2.values(), columns = channels).mean(axis = 0)
APA_synth_od_PCA_3 = pd.DataFrame(APA_synth_od_PCAs_3.values(), columns = channels).mean(axis = 0)
APA_synth_od_ICA = pd.DataFrame(APA_synth_od_ICAs.values(), columns = channels).mean(axis = 0)
APA_synth_od_tddr = pd.DataFrame(APA_synth_od_tddrs.values(), columns = channels).mean(axis = 0)
APA_synth_od_cbsi = pd.DataFrame(APA_synth_od_cbsis.values(), columns = channels).mean(axis = 0)
APA = pd.concat([APA_synth_od_WF, APA_synth_od_kWF, APA_synth_od_PCA_2, APA_synth_od_PCA_3, APA_synth_od_ICA, APA_synth_od_tddr, APA_synth_od_cbsi], axis = 1)
APA.columns = methods

APA.plot.bar(title = 'APA scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'APA (dB)', xlabel = 'Channels').get_figure().savefig('APA.png', bbox_inches = 'tight')

# Mean score over all channels
scores_APA_synth = APA.mean(axis = 0).sort_values(ascending = False)
scores_APA_synth

cbsi     475.902631
WF       251.010353
kbWF      67.060958
ICA       47.092512
tddr       0.558196
PCA_3      0.000000
PCA_2    -14.239547
dtype: float64

print('The best MA correction technique based on APA value is: ' + APA.mean(axis = 0).idxmax() + ' (value = ' + str(APA.mean(axis = 0).max()) + ')')

The best MA correction technique based on APA value is: cbsi (value = 475.90263122389865)

channels = raw_od.ch_names

APA2 = pd.DataFrame()
APA2['WF'] = APA_od_WF
APA2['kbWF'] = APA_od_kWF
APA2['PCA_2'] = APA_od_PCA_2
APA2['PCA_3'] = APA_od_PCA_3
APA2['ICA'] = APA_od_ICA_MSD
APA2['ICA_CEI'] = APA_od_ICA_CEI
APA2['tddr'] = APA_od_tddr
APA2['cbsi'] = APA_od_cbsi
APA2.index = channels

%matplotlib qt
APA2.plot.bar(width = 1.5, title = 'APA scores per channel and MA correction method (Averaged over 10 synthetic datasets)', ylabel = 'APA (dB)', xlabel = 'Channels', figsize = (30,5)).get_figure().savefig('APA2.png', bbox_inches = 'tight')

# Mean score over all channels
scores_APA = APA2.mean(axis = 0).sort_values(ascending = False)
scores_APA

cbsi       186.927048
WF         152.948228
kbWF        50.150134
ICA         14.147645
tddr         0.018830
PCA_3        0.000000
PCA_2      -17.451236
ICA_CEI    -21.499368
dtype: float64

print('The best MA correction technique based on APA value is: ' + APA2.mean(axis = 0).idxmax() + ' (value = ' + str(APA2.mean(axis = 0).max()) + ')')

The best MA correction technique based on APA value is: cbsi (value = 186.92704761292265)

results = {
    'SNR': scores_SNR,
    'PRD': scores_PRD,
    'CC' : scores_CC,
    'CNR_Cooper' : scores_CNR_C,
    'CNR_Zhang' : scores_CNR_Z,
    'APA_synth' : scores_APA_synth,
    'APA' : scores_APA.drop('ICA_CEI')
}
ranking = pd.DataFrame()
# Calculate ranks for each metric
for metric_name, metric_scores in results.items():
    # Rank the scores; ascending or descending depending on the metric's nature
    if metric_name == 'PRD':  
        ranking[f'{metric_name}_rank'] = metric_scores.rank(ascending=False, method='min').astype(int)
    else:  
        ranking[f'{metric_name}_rank'] = metric_scores.rank(ascending=True, method='min').astype(int)

# Display the final DataFrame
ranking

if export:
    dfi.export(ranking,"ranking_MA_methods_table.png")

%matplotlib inline
# Sum rankings
final_ranking = ranking.sum(axis = 1).sort_values(ascending = False)
final_ranking.plot.bar(title = 'Global ranking of MA correction alorithmns', ylabel = 'Summed ranking of all evaluation metrics', xlabel = 'Motion artefact correction methods', figsize = (10,5)).get_figure().savefig('ranking_methods.png', bbox_inches = 'tight')

Participant
Measurement date	May 08, 2024 13:35:00 GMT
Experimenter	Unknown

Digitized points	20 points
Good channels	52 fNIRS (CW amplitude)
Bad channels	None
EOG channels	Not available
ECG channels	Not available

Sampling frequency	5.88 Hz
Highpass	0.00 Hz
Lowpass	2.94 Hz
Filenames	MA_experiment_data.snirf
Duration	00:14:38 (HH:MM:SS)

Participant
Measurement date	Unknown
Experimenter	Unknown

Digitized points	Not available
Good channels	4 Oxyhemoglobin, 4 Deoxyhemoglobin
Bad channels	None
EOG channels	Not available
ECG channels	Not available

	onset	duration	description
0	2024-05-08 13:36:04.600	3.0	practice/Horizontal
1	2024-05-08 13:36:25.170	3.0	practice/Random
2	2024-05-08 13:36:46.760	3.0	practice/Vertical
3	2024-05-08 13:37:08.180	3.0	practice/Vertical
4	2024-05-08 13:37:28.240	3.0	practice/Horizontal
5	2024-05-08 13:37:46.430	3.0	practice/Random
6	2024-05-08 13:38:21.000	3.0	Baseline
7	2024-05-08 13:38:25.000	3.0	Baseline
8	2024-05-08 13:38:29.000	3.0	Baseline
9	2024-05-08 13:38:34.000	3.0	Baseline
10	2024-05-08 13:38:39.000	3.0	Baseline
11	2024-05-08 13:38:42.000	3.0	Baseline
12	2024-05-08 13:38:43.000	3.0	Baseline
13	2024-05-08 13:38:59.000	3.0	Baseline
14	2024-05-08 13:39:04.000	3.0	Baseline
15	2024-05-08 13:39:10.000	3.0	Baseline
16	2024-05-08 13:40:11.100	3.0	test/Random
17	2024-05-08 13:40:32.520	3.0	test/Vertical
18	2024-05-08 13:40:51.560	3.0	test/Random
19	2024-05-08 13:41:11.280	3.0	test/Vertical
20	2024-05-08 13:41:27.770	3.0	test/Horizontal
21	2024-05-08 13:41:45.960	3.0	test/Horizontal
22	2024-05-08 13:42:02.280	3.0	test/Random
23	2024-05-08 13:42:22.170	3.0	test/Vertical
24	2024-05-08 13:42:41.890	3.0	test/Random
25	2024-05-08 13:42:59.400	3.0	test/Random
26	2024-05-08 13:43:18.270	3.0	test/Random
27	2024-05-08 13:43:35.610	3.0	test/Horizontal
28	2024-05-08 13:43:52.100	3.0	test/Vertical
29	2024-05-08 13:44:10.290	3.0	test/Horizontal
30	2024-05-08 13:44:28.480	3.0	test/Vertical
31	2024-05-08 13:44:45.650	3.0	test/Vertical
32	2024-05-08 13:45:06.390	3.0	test/Vertical
33	2024-05-08 13:45:27.810	3.0	test/Horizontal
34	2024-05-08 13:45:46.000	3.0	test/Horizontal
35	2024-05-08 13:46:02.490	3.0	test/Vertical
36	2024-05-08 13:46:20.510	3.0	test/Random
37	2024-05-08 13:46:37.170	3.0	test/Random
38	2024-05-08 13:46:53.660	3.0	test/Vertical
39	2024-05-08 13:47:15.080	3.0	test/Horizontal
40	2024-05-08 13:47:33.270	3.0	test/Random
41	2024-05-08 13:47:52.990	3.0	test/Horizontal
42	2024-05-08 13:48:11.350	3.0	test/Vertical
43	2024-05-08 13:48:31.920	3.0	test/Horizontal
44	2024-05-08 13:48:49.260	3.0	test/Random
45	2024-05-08 13:49:06.600	3.0	test/Horizontal

	SNR_rank	PRD_rank	CC_rank	CNR_Cooper_rank	CNR_Zhang_rank	APA_synth_rank	APA_rank
tddr	7	7	7	5	4	3	3
cbsi	6	6	1	7	2	7	7
ICA	5	5	6	6	3	4	4
WF	4	4	3	2	5	6	6
kbWF	3	2	5	4	1	5	5
PCA_3	2	3	2	1	7	2	2
PCA_2	1	1	4	3	6	1	1

Motion artefact removal¶

Table of Contents¶

Pre-processing ¶

Load python modules ¶

Import data ¶

Import snirf data ¶

Add info to raw intensity data ¶

Adress problem with timing of fNIRS recordings¶

Signal quality metrics and channel inspection ¶

Validate that the location of sources-detector pairs and channels are in the expected locations¶

Visual inspection of the raw light intensity data¶

Visual inspection of the the optical density¶

Visual inspection of the haemoglobin concentration¶

Scalp-coupling index¶

SCI over whole signal¶

SCI evaluated over moving window¶

Peak spectral power¶

Import IMU data ¶

Create synthetic fNIRS data ¶

Create experimental design ¶

Introduce serial correlations by means of AR model ¶

Introduce physiological confound ¶

Add synthetic HRFs ¶

Introduce motion artefacts ¶

Convert back to OD (took me a while to figure out)¶

Inspect signal quality ¶

Visual inspection (example on first set)¶

SCI¶

PSP¶

Convert to haemodynamic concentration¶

Create evaluation metrics ¶

Metrics requiring ground truth ¶

Signal-to-noise ratio¶

Percent root difference¶

Correlation coefficient¶

Metrics not requiring ground truth ¶

Contrast-to-noise ratio¶

Artifact power attenuation¶

Auxiliary hardware algorithms - Not finished¶

Accelerometer-based motion artifact removal algorithm (NOT WORKING) ¶

Active noise cancelation ¶

Signal processing methods ¶

Wavelet filtering ¶

Perform on synthetic data¶

Evaluate filter method¶

Perform on motion artefact experiment data¶

Evaluate filter method¶

Kurtosis-based wavelet filtering ¶

Perform on synthetic data¶

Evaluate filter method¶

Perform on motion artefact experiment data¶

Evaluate filter method¶

Spline interpolation (NOT WORKING) ¶

Perform on synthetic data¶

Principal component analysis ¶

Perform on synthetic data¶

Examine PCA components¶

Evaluate filter method (Only stimulus segement)¶

Perform on motion artefact experiment data¶

Evaluation filter method¶

Independent component analysis ¶

Perform on synthetic data¶

Evaluate filter method¶

Perform on motion artefact experiment data¶

Evaluate filter method¶

Temporal derivative distribution repair

Perform on synthetic data¶

Evaluate filter method¶

Perform on motion artefact experiment data¶

Evaluate filter method¶

Correlation-based signal improvement ¶

Perform on synthetic data¶

Evaluate filter method¶

Perform on motion artefact experiment data¶

Evaluate filter method¶

Compare motion artefact techniques ¶

Synthetic datasets ¶

SNR¶

PRD¶

DeltaCC¶

Auxiliary hardware algorithms - Not finished ¶